July 2020

Advanced Technologies
for Industry – AT WATCH
Technology Focus on Artificial Intelligence
 AT WATCH- European Commission

This report was prepared by Gabriella Cattaneo, Andrea Siviero, Giorgio Micheletti, Giulia Carosella
and Irene Magnani (IDC).

EUROPEAN COMMISSION
Executive Agency for Small and Medium-sized Enterprises (EASME)
Unit A.1.2 — COSME
E-mail: EASME-COSME-ENQUIRIES@ec.europa.eu
Directorate General for Internal Market, Industry, Entrepreneurship and SMEs
Unit F.1 — Industrial Strategy and Value Chains
Unit F.2 — Social Economy
E-mail: GROW-ATI@ec.europa.eu

European Commission
B-1049 Brussels

LEGAL NOTICE
The information and views set out in this report are those of the author(s) and do not necessarily reflect the official
opinion of EASME or of the Commission. Neither EASME, nor the Commission can guarantee the accuracy of the data
included in this study. Neither EASME, nor the Commission or any person acting on their behalf may be held
responsible for the use, which may be made of the information contained therein.
More information on the European Union is available on the Internet (http://www.europa.eu).

Print ISBN 978-92-9202-960-9 doi:10.2826/108632 EA-04-20-275-EN-C

PDF ISBN 978-92-9202-961-6 doi:10.2826/004152 EA-04-20-275-EN-N

© European Union, 2020

July 2020
 AT WATCH- European Commission

Table of contents

Table of contents ................................................................................................................... 3

Section 1 ................................................................................................................................ 4
1. Introduction ................................................................................................................. 4
1.1 The Advanced Technologies Industry landscape .................................................................... 5
Section 2 .............................................................................................................................. 10
2. Technology Focus: Artificial Intelligence ...................................................................... 10
2.1 The Market Potential .............................................................................................................. 11
2.1.1 Adoption and industry trends ....................................................................................................................................... 11
2.1.2 Main Barriers and Challenges................................................................................................................................... 13

2.2 Use cases and business opportunities ................................................................................. 15

2.2.1 AI value proposition.................................................................................................................................................. 15
2.2.2 Leading AI Use Cases ................................................................................................................................................ 15

2.3 Social and Sustainability Impacts ....................................................................................... 18

2.3.1 Introduction .............................................................................................................................................................. 18
2.3.2 AI for social good ...................................................................................................................................................... 19
2.3.3 AI in the Healthcare Sector ....................................................................................................................................... 22

2.4 Policy, Regulatory and Ethical Implications ......................................................................... 25

2.4.1 Recent Policy Developments: An Overview ............................................................................................................... 25
2.4.2 AI and the need of trustworthiness .......................................................................................................................... 26

2.5 Conclusions........................................................................................................................ 28
Appendix A: Bibliography ..................................................................................................... 30
Appendix B: Technology definitions ...................................................................................... 33
Appendix C: Advanced Technology uptake ........................................................................... 36
About the ‘Advanced Technologies for Industry’ project ....................................................... 37

July 2020

3
 AT WATCH- European Commission

Section 1

1. Introduction

This Advanced Technology Watch report has been developed in the framework of the ‘Advanced
Technologies for Industry’ (ATI) project, initiated by the European Commission, Directorate General for
Internal Market, Industry, Entrepreneurship and SMEs and the Executive Agency for Small and Medium-
sized Enterprises.
As part of a series of analytical reports on advanced technologies’ trends, this report represents the first
issue of a comprehensive monitoring tool endowing policymakers, industry, researchers, and other
relevant stakeholders with regularly updatable research. The AT-Watch report series is meant to play a
complementary role to the other analytical, policy and statistical reports of the project, by focusing on
market, business, and socio-economic trends driven by technology innovation. This Advanced
Technology Watch encompasses therefore the whole set of advanced technologies that are a priority for
the European industrial policy. These technologies enable process, product, and service innovation
throughout the economy, thus fostering industrial modernisation.
The qualitative and quantitative analysis included in this Advanced Technology Watch is specifically
designed to provide novel insight and up-to-date content to technology users across the whole set of
the European industry. This, with the aim to unveil the potential opportunities emerging from the most
recent applications of advanced technologies.
The AT Watch report series aims at:
• a primary audience of industry stakeholders, including SMEs, and industry associations
interested to learn about upcoming technology trends and business opportunities;
• a complementary audience of national/regional/local policy makers, interested to support
industry in the exploitation of technology innovation and emerging business opportunities, by
removing barriers and creating favourable market conditions;
• a complementary audience of research and technology stakeholders interested in the applied
research challenges to be solved in order to capture the emerging business opportunities.

Each report is thus structured along two main sections:

• A brief overview of all Advanced Technologies uptake and demand trends by industry (Section
1: Technology Landscape)
• A more in-depth analysis of one of the Advanced Technologies, selected because of its relevance
in terms of emerging business opportunities and disruptive potential (Section 2: Technology
Focus).
In this report, section 2 is devoted to Artificial Intelligence (AI) as one the most significant technologies
that are currently shaping the whole process of digital transformation and industrial modernisation in
Europe and Worldwide.

July 2020

4
AT WATCH- European Commission

1.1 The Advanced Technologies Industry landscape

The digitisation and industry modernisation process in Europe is progressing at different speeds
across all industry sectors, driven by a whole set of changing priorities, challenges and use cases. In
such a dynamic scenario, advanced technologies’ deployment may take different forms across European
industries, depending on the levels of their technological maturity, the availability and relevance of
industry-specific applications and the expected and desired business, operational and societal outcomes
that a sector intends to achieve.
The different mix of advanced technologies adopted by each industry is visualised in Figure 1 below.
The figure shows the percentage share of enterprises in each industry adopting or planning to adopt
each technology (the size of bubbles corresponds to the level of uptake, with the highest value being
88%). The data is based on the Advanced Technologies for Industry Survey (July 2019) 1 and on a
sample of European enterprises from 11 Member States, representing altogether 85% of the EU GDP in
2019.
Figure 1: Advanced Technologies Uptake by European Industries, 2019

Source: Advanced Technologies for Industry Survey July 2019, (N=900). Legend: Bubble size represents the % of
enterprises in the industry adopting the technology in the same row. The maximum value is 88%. Appendix B
provides the definition of each technology.

The visualisation highlights how a distinct group of technologies features a marked horizontal
diffusion across all industries (general purpose technologies including public cloud, Security, mobile
solutions, Big Data, Internet of Things (IoT), Advanced Connectivity and Artificial Intelligence): they
represent the technology portfolio necessary (but not sufficient) for digital transformation. Other
technologies clearly display a niche or industry-specific orientation. However, this does not
mean that they do not provide opportunities for investments outside their main industry niche. Taking
Blockchain as an example, the technology initially found its ground in financial applications. New areas
of applications are now emerging, and multiple novel use cases proliferate in order to drive business
value in other industries. For instance, in manufacturing, Blockchain is used to keep track and certify
product sources along the value chain. Similarly, Blockchain exhibits a great potential also in the art
market where it can be used not only for tokenisation of artworks sales, but also to verify the provenance
and authenticity of artworks - one of the biggest challenges for the industry.

1The survey interviewed a sample of 900 enterprises with more than 10 employees in CZ, DK, DE, FR, ES, HU, IT, NL, PL,
SE, UK.

July 2020

5
AT WATCH- European Commission

When looking at the European industries in more detail, we observe that:

• the manufacturing industry has a strong focus on the development of an Industry 4.0
strategy prioritising technologies that can provide clear benefits in terms of driving operational
performance and reducing costs. Automation is a strong focus and will help companies to
simplify complex tasks or processes and relieve human resources of heavy and time-consuming
workloads. An important role is played by Robotics or collaborative robots (cobots) that help
manufacturers achieve efficiency benefits and enable staff to save time. Efficiency gains are also
enabled by Augmented Reality / Virtual Reality (ARVR) solutions that allow experts to
provide remote support to on-field operators and guide them through step-by-step instructions.
R&D and product innovation is another strong manufacturing priority driving experimentation
with several emerging technologies. In this context, AI and Big Data/analytics are helping
manufacturers to improve how they design, manufacture and deliver their products. 3D
printing is also poised to grow as it will automate ways to create prototypes or new product
parts, reducing production times and improving products. Product innovation is also driving
adoption of Advanced Materials, micro and nano electronics, nanotechnologies and photonics
with the aim to improve products and reduce costs.
• In finance, besides operational efficiency, the other main business goal driving investments in
advanced technologies is the need to attract and retain customers. This is pushing the industry
towards piloting new service delivery models. AI, Big Data and Blockchain are the most
promising technologies for the industry as they enable automation of internal operations,
improved customer service and greater protection against security threats. Key AI trends in the
industry include automation of IT operations and opening new digital channels to improve
customer experience leveraging voice banking and chatbots. Blockchain main applications
include for example cross-border payments and settlements. The industry has also been central
to the emergence of a new digital economy, the open banking. The open banking digital
economy is connected to the European payment services directive (PSD22). The core of this
initiative is the request for banks to share more customer information than ever before via
technologies such as application programming interfaces (APIs).
• For telecom and media providers new technologies and new customers behaviours are
generating several opportunities to boost current income sources and generate new revenue
streams. 5G technologies for example are expected to provide many monetisation
opportunities for telco unleashing the potential of other technologies such as ARVR and IoT.
Streamlining processes, improving network efficiency and improving customer support will be
the main drivers of telco’s digital investments. Interesting pockets of growth can be found also
in investment in other advanced technologies, such as Photonics that are supporting the
development of fibre optic network communications. Media transformation process had been
driven by changing customer needs and behaviours. New channels and platforms for distributing
and accessing content and new ways for producing content are driving innovation in the
industry. Streaming, content-as-a service and new technologies for creating engaging
entertainment experiences, such as wearables and ARVR technologies, are major trends under
the spotlight.
• Utilities and Oil&Gas show interesting opportunities in terms of many advanced technologies.
Hot spots in the industry are the e-mobility revolution and AI powered home energy
management. Electric vehicles are expected to be a mass-market revolution, driven by
increasing sustainability concerns and blurring industry boundaries between power distribution
and retail, transportation and automotive. Large and small utilities can play different roles across
the entire value chain from electric vehicle supply equipment, to management platforms, to
meter and grid services, to home and in-network charging. The quest for alternative and
sustainable energy sources is also paving the way to the use of advanced technologies such
as photonics and nanotechnologies for power generation and for new and more efficient
lighting solutions. The second growing trend is the smart home ecosystem where utilities can

2
https://ec.europa.eu/info/law/payment-services-psd-2-directive-eu-2015-2366_en

July 2020

6
AT WATCH- European Commission

play a big role in providing advanced home energy management solutions and automation
functionalities using devices such as smart plugs, thermostats and smart lighting for optimising
energy consumption but also for getting insights into consumers' habits.
• The healthcare industry shows some interesting investments in Biotechnology, ARVR,
Nanotechnology, Advanced Materials and Photonics, compared to the other industries.
Uptake of advanced technology in the industry is strictly linked to the need to innovate and
improve patient care, providing integrated and personalised services. There is a strong need for
data sharing between doctors, different medical units or hospitals to enable faster services and
avoid duplicating efforts. Higher investments in cloud suggest interest toward infrastructure
and operations optimisation. Robots, especially for surgery and logistics purposes, are
becoming more affordable, and hospitals will start to invest more significantly in the upcoming
years. Investments in wearables, IoT and AI are growing with the need to monitor patient
behaviour and accidents for elderly people with medical conditions to provide prompt emergency
help. Devices such as ARVR devices are helping doctors improve surgery and diagnosis and are
also used for therapeutic purposes (e.g. rehabilitation).
• Due to the highly customer-centric nature of retail, efforts in providing superior and
differentiated customer experience are key in this industry. The ecommerce channel, where
consumers can finalise their purchases using their PCs or mobile phones, remains a priority for
retailers and a successful strategy during the COVID-19 pandemic crisis. As more customers
are switching to mobile commerce, customer assistance and support are also changing. Through
AI-enabled chatbots, customers can contact companies on social platforms to track
shipments, request product refunds or raise complaints. Through deep learning as well as text
and voice recognition, AI investments will be heavy across the distribution sector, and they will
focus on improving customer experience. A growing opportunity in the industry is represented
by real-time contextual personalisation for the customer, which allows retailers to shape
the customer experience in relation to multiple parameters such as demographics, location,
day/time, weather and purchasing patterns.
• The pattern of technology adoption by Government and Education is influenced by the
national context. The investments in innovation by public administrations mirror the budget
availability at Member State or regional level. Governments are working to streamline internal
bureaucratic processes through automation. In fact, long back-office tasks slow down critical
government work, resulting in longer waiting times for citizens to access services. AI solutions
based on facial recognition are also seeing significant investment, for example to help police
recognise and identify criminals. Smart city projects, combining mobile, IoT and Big
Data/Analytics solutions, are expected to push investments in technology, especially for safety
purposes (such as video surveillance) and for public transport optimisation. Another driving
trend in the industry is represented by the open data portals, with the aim to improve
transparency, openness and interaction by sharing public data with citizens. Education
institutions in Europe are prioritising investments in mobile solutions, for example investing
in the provision of mobile devices. Schools are also showing an increasing interest in distance
learning, with the development of online platforms and e-learning apps for students. After the
emergency-driven experience of distance learning during the lockdown period, forced by the
COVID-19 pandemic, it is possible that this trend will accelerate as many schools have
appreciated their potential advantages (and acknowledged their disadvantages). Many
governments and schools are likely to increase investments in distance learning, in a
complementary role to in presence education.
• Client expectations, including value-for-money, fast delivery times and high quality of work,
together with increasing competition, are putting a lot of pressure on the Professional
Services industry. Therefore, the industry will continue to invest in advanced technologies to
add value to services and business models. Digital technologies are changing the industry in
their client-facing and back-end activities. For example, advanced technologies will be able to

July 2020

7
AT WATCH- European Commission

automatically process documents such as legal, shareholder and market reports, impacting
positively on timing and freeing staff from tasks that can be automated. Cloud Computing
continues to attract interest, mainly driven by flexibility that the solution offers in terms of
accessing information anytime and increasing collaboration between teams and the low cost to
deploy the technology. Professional services are a data intensive vertical investment in security
and will be driven by the implementation of GDPR (General Data Protection Regulation). This
will drive the industry to raise technology barriers to protect client sensitive information and
avoid data breaches.
• The uptake by the Transport industry of advanced technologies is customer and data-
centric, aiming at delivering enhanced experience to customers while leveraging data and
analytics to optimise operations and streamline processes. IoT and AI play a key role in
supporting industry companies to regulate traffic flows, streamline security checkpoints with
biometrics such as facial recognition, and reduce the number of lost bags using electronic
luggage tags. In logistics, heavy workloads can be eased by introducing solutions to create
collaborative environments in which humans coexist with robots, with the latter taking over
heavy, repetitive and time-consuming tasks. Smart trucks3 are gaining popularity due to the
ability of automating processes through the integration of different types of technologies and
enabling real-time emergency and incident reactions.
• The agriculture sector faces many challenges, including increasing demand for food, climate
change and lack of workers. Advanced technologies are increasingly playing a role in responding
to some of these challenges. Data-driven innovation is transforming farm management
through the so-called precision agriculture approach. By leveraging satellites, drones and
IoT sensors in farm equipment (such as tractors), an unprecedented amount of data can be
collected to monitor the conditions of the crops, soil and other key elements for cultivation, as
well as cattle. This allows farmers to manage irrigation, fertilisation and all the farming
processes in a scientific way, minimising costs and the use of pesticides and maximising outputs.
The digitisation of the farming processes represents also the first step for the emerging food
track-and-tracing systems developed to guarantee quality and safety, highly appreciated in the
food-agriculture value chain. This represents a truly digital transformation for a sector which in
the past used very little ICT. This innovation process is profoundly changing the culture,
organisational processes and skills requirements in the sector, facing barriers of reluctance and
difficulties to adapt. Advanced technologies show interesting uses also in fighting climate
change and related risks (such as the loss of arable land and increased urbanisation). For
example, a growing trend is represented by urban or vertical farms, leveraging technologies
to minimise the use of natural resources such as soil, water and energy. This is done by using
IoT and photonics to manage parameters such as humidity, light and irrigation to get the most
out of crops. Industrial biotechnology shows also interesting promises for obtaining
alternative healthy, protein-rich and nutritionally balanced food raw material responding to
increasing population and food demand.

This overall picture of advanced technologies’ deployment in the European industry is to a certain extent
reflected by the analysis of the demand and supply of advanced technologies’ skills carried out in the
report on the General Findings within the framework of the present project4. In terms of skills supply,
and based on the profile of registered users on LinkedIn, the share of advanced technology skilled
professionals (vis-à-vis the total number of professionals) in selected industries reveals that Europe’s
manufacturing industry absorbs the highest number of skilled professionals. This is particularly true for
the Automotive sector where technologies such as Advanced Manufacturing and IoT are clearly
instrumental for the development of Industry 4.0 strategies.
Other industries such as Electronics and, to a lesser extent, Chemicals, employ a large amount of skilled
professionals, especially for technologies like Advanced Manufacturing and IoT (in Electronics) and

3 The adjective “Smart” refers to multiple tech-enabled scenarios making trucks intelligent, such as e-call equipped trucks
(i.e. calling automatically emergency in case of crash), driver fatigue monitoring (systems that track drivers’ attention level
and notify in case of risk).
4 ATI General findings (D3.4), Section 5, June 2020

July 2020

8
AT WATCH- European Commission

Advanced Materials and Industrial Biotech (in Chemicals), confirming that manufacturing as a whole
remains at the forefront of the the digital transformation and modernisation processes in Europe.
In terms of skills demand, manufacturing exhibits high levels of hiring positions measured by the number
of online job advertisement requiring specific skills. Again, the Automotive sector requires specific skills
in Advanced Manufacturing, AI and Robotics, just like the Electrical & Electronics exhibits strong demand
of skills in Advanced Materials, Micro-nanoelectronics, nanotechnologies and Robotics. The prominence
of the Manufacturing industry is challenged only by the Finance sector where, both in Banking and
Financial Services, specialised skills for Big Data, Blockchain, Cloud Computing and cybersecurity are
very much in demand across Europe.

July 2020

9
 AT WATCH- European Commission

Section 2

2. Technology Focus: Artificial Intelligence

Artificial Intelligence (AI) has fascinated human imagination since 1950, when Alan Turing first posed
the question of whether machines can think, but until recent years it never delivered on its promises.
Since 2011, breakthroughs in machine learning (ML), a technique leveraging a statistical approach to
make predictions from historical data, and neural networks (a modelling technique) have completely
changed the AI technology environment. The availability of huge datasets and technology advances in
Big Data, the Internet of Things (IoT) and fast connectivity are the enablers of AI new
systems and services, digital assistants, robots and drones. Even without an intelligence truly
comparable to the human one, the achievements of AI in complementing and substituting intelligent
human activities are impressive. The pervasiveness of AI and its disruptive potential of value chains and
business models make it both an opportunity and a threat for European enterprises’ competitiveness.
Today’s AI is a general-purpose technology, a collection of tools that can be combined, extended and
packaged into different types of applications across all industries, such as voice and image recognition,
machine translation, control of assisted driving and autonomous vehicle navigation. Given this
multidimensionality, there are multiple definitions of AI.
The EC’s White Paper on artificial intelligence5, published for public consultation in February 2020,
defines AI as “a collection of technologies that combine data, algorithms and computing
power” and one of the most important applications of the data economy. The EC considers AI as an
instrument to harness the value of data in order to build European sustainable economic growth and
societal wellbeing. A trustworthy AI is an essential building block of Europe’s ambition to build a digital
society on European values and rules6.
The OECD AI Expert Group (AIGO)7 defines AI more narrowly as “machine-based systems that can, for
a given set of human defined objectives, make predictions, recommendations, or decisions influencing
real or virtual environments”.8
This report is focused on a definition, which is coherent with both the EC and the OECD definitions and
is focused on AI’s capability to support decision making, classifying AI “as systems that learn, reason,
and self-correct. These systems hypothesise and formulate possible answers based on available
evidence, can be trained through the ingestion of vast amounts of content, and automatically adapt and
learn from their mistakes and failures”9.
Recommendations, predictions, and advice based on this AI framework provide users with answers and
assistance in a wide range of applications and use cases. The operational definition in focus specifies
that AI software technologies include natural language processing (NLP), image and video
analytics, machine learning (ML), knowledge graphs and other technologies to answer questions,
discover insights and provide recommendations. AI also empowers innovation in Robotic Process
Automation (RPA: software solutions powering robots) and drones.
The new generation of AI technologies entering the market is far from mature. Its full exploitation will
require deep transformation of organisational and business processes as well as skill changes and new
ways of working. Digital transformation remains the key to the successful adoption of these
technologies. As for every new technology, AI will need to be “domesticated” by the socio-economic
system to reach its full potential. Issues such as trust, transparency, ethical and social challenges will
need to be addressed, as discussed in Section 2.4 “Policy, Regulatory and ethical implications”.
Unsurprisingly, the advent of AI plays a pivotal role in the EU and Member States’ policy strategies.
Alarmed by a gap in R&D investments in innovative technologies compared with China and the US, the
EU launched an ambitious strategy already in 201810.

5 European Commission “White Paper on Artificial Intelligence - A European approach to excellence and trust”
6 European Commission “Shaping Europe’s Digital Future”, February 2020
7 OECD (2019), "Scoping the OECD AI principles: Deliberations of the Expert Group on Artificial Intelligence at the OECD

(AIGO)",
8 OECD, “Recommendation of the Council on Artificial Intelligence” (2019)
9 IDC’ Worldwide Artificial Intelligence Taxonomy, 2019
10 EC Communications “Artificial Intelligence for Europe” and “Coordinated Plan on Artificial Intelligence”, 2018

July 2020

10
AT WATCH- European Commission

The White Paper on Artificial Intelligence11 published in February 2020 builds on these previous
steps with renewed ambition and investments to build cutting-edge digital capabilities in the field. Two
main strands of policy are planned. The first aims at building an AI European ecosystem of excellence,
leveraging the forthcoming Framework Programs Horizon Europe (HE) and Digital Europe (DE) in the
period 2021-2027, including a new public-private partnership in AI and Robotics. The plan includes
investments in AI research excellence centres and AI Digital Innovation hubs, as well as increased equity
funding through the European Investment Fund. The second policy strand aims at building an ecosystem
of trust, developing a legislative framework favourable to the development of AI, dealing with AI-related
risks, guaranteeing the respect of human rights and European values. In fact, concerns about the social
and ethical risks from the development of AI have given raise to multiple guidelines from international
organisations and governments, as illustrated in Section 2.4.1 “Recent policy developments”.
To help industry stakeholders, national/regional/local policy makers, as well as researchers and
technology stakeholders to obtain an up-to-date picture of AI’s latest developments in Europe, the
report will now focus on the market potential associated with AI. It will further investigate AI value
proposition in terms of leading use cases of this technology in Europe, in order to highlight the concrete
business opportunities stemming from the adoption of the AI. Finally, the social dimension and impacts
of AI on the society as whole, together with their policy-related and ethical implications, will be
considered in Section 2.4 .

2.1 The Market Potential

2.1.1 Adoption and industry trends
After an early stage dominated by innovative players and exploratory applications in specific business
and societal scenarios, artificial intelligence has rapidly gained in importance and diffusion and is now
beyond the nascent phase. The potential for future growth and economic impacts is still impressive,
because AI is a general-purpose technology expected to be adopted in different ways by all industries
and all social actors. According to Mc Kinsey estimates presented in 2019 12 “If Europe develops and
diffuses AI according to its current assets and digital position relative to the world, it could add some
€2.7 tn, or 19 percent, to output by 2030. Such an impact would be roughly double that of other general-
purpose technologies adopted by developed countries in the past”. Even if these estimates need to be
revised because of the disruption caused by the COVID-19 pandemic, there is still no doubt about the
relevance of the potential impacts of AI-driven innovation. Other sources, such as the OECD 13, confirm
that artificial intelligence is poised to span every aspect of our daily lives.
The speed of evolution of the market is impressive. At worldwide level, IDC forecasts 14 the total
spending on AI to grow at a 5-years Compound Annual Growth Rate (CAGR) of 26.5% for the period
2018–2023, up to €96 bn at the end of that period. The European Union market is expected to grow
faster than the global market, representing a share of 23% of total by 2023 (versus 18% in 2018). An
overview of the current AI spending in Europe in terms of percentage shares across the different
industries in offered in Figure 2 below.

11 European Commission Communication, “White Paper on Artificial Intelligence - A European approach to excellence and
trust” 2020
12 Mc Kinsey Global Institute, Notes from the AI Frontier: “Tackling Europe’s gap in digital and AI”, 2019
13 OECD (2019) “Artificial Intelligence in Society”
14 IDC Worldwide Artificial Spending Guide, March 2020

July 2020

11
AT WATCH- European Commission

Figure 2: Total AI spending % value by Industry, Europe, 2019

Total Europe AI Spending by Industry in 2019

7%
Financial Services 7% 24%
Manufacturing
5%
Retail and Wholesale
4%
Government
Telecom and Media
10%
Healthcare
Transport Services 22%
7%
Professional Services
Others 14%

Source: IDC Worldwide Artificial Intelligence Spending Guide, March 202015

In Europe, the use of AI is growing as existing users deepen their investments and new users enter the
market. According to the Advanced Technologies for Industry Survey (July 2019), more than one fourth
of European organisations are currently adopting AI and another half are planning or evaluating to adopt
AI systems in the near term. Figure 3 below displays the AI adoption rates across European industries
according to the results of the Advanced Technologies for Industry Survey (July 2019).
Sectors such as Finance, Telecom/media and Utilities/Oil & Gas take the front seat in the European
AI scenario. Having been trailblazers since the first years of AI experiments and proof-of-concepts, they
continue to lead the way in terms of AI adoption, driven by their data-intensive business nature that
provides a fertile ground to AI applications. For example, one out of two European financial companies
are already adopting AI technologies16, enhancing their workforce skills with AI tools that provide better
analysis and investigation capabilities.

15 Total Europe (as per IDC definition):

AL, AT, BE, BA, BGR, HR, CZ, DK, EE, FI, FR, DE, GR, HU, IRL, IT, KAZ, LVA, LT, MK, MNE, NL, NO, PL, PT, RO, RU, RS, SK,
SI, ES, SE, CH, UA, UK and Rest of CEE.
16 Source: Advanced Technologies for Industry Survey July 2019

July 2020

12
AT WATCH- European Commission

Figure 3: AI Adoption across European Industries (share of respondents)

Source: Advanced Technologies for Industry Survey July 2019, (N=900)

Both Discrete and Process Manufacturing show a percentage of respondents planning to adopt AI this
year (25.4% and 23.1%, respectively) above total European average (22.8%) (Figure 3), driven by AI
capability to support end-to-end value chain visibility and implement data-driven processes in factories.
Even if the 2020 recession due to the COVID-19 pandemic will slow down technology adoption plans in
the short term, AI investments are expected to start growing again soon in the post-crisis scenario, to
fuel recovery through digital transformation and increased automation.
2.1.2 Main Barriers and Challenges
The actual deployment of AI use cases is hindered by multiple challenges and implementation barriers.
To reap concrete business opportunities, enterprises must be aware of these challenges and of the
countermeasures that industry and policy makers are trying to implement in Europe. Enterprises should
also face these challenges in a proactive way, without waiting for them to be solved, to avoid losing the
competitive AI race.
The main challenges standing in the way of a full-blown AI development and deployment in Europe are:

• The need to strengthen the AI research and innovation environment, which is

fragmented17. This creates difficulties to bring competitive innovations to the global market. The
AI Digital Innovation Hubs to be created across Europe have the objective to help bridge this
gap between research and industry. Building essential digital capabilities in Europe for AI is one
of the strategic objectives of the forthcoming Digital Europe Programme18.
• Increasing demand for skills and know-how. The number of tech-savvy professionals does
not meet the exponentially increasing current demand. Many European organisations do not
have the skills needed to deploy and manage AI solutions19. The hiring demand for AI is
particularly high in the Financial services and Automotive sectors.20 Developing advanced digital
skills for AI as well as other strategic technologies is one of the missions of the forthcoming
Digital Europe Programme, as the EC has identified the lack of skills as one of the critical
bottlenecks of European digital innovation capability.
• An insufficient availability of data and data sharing. Datasets are the lifeblood of Artificial
Intelligence. Global platforms such as Google and Facebook dominate the access to personal

17 “European Artificial Intelligence. (AI) leadership, the path for an integrated vision”. Policy Department for Economic,
Scientific and Quality of Life Policies, Directorate-General for Internal Policies. Laura Del Ponte (CSIL) 2018
18 https://ec.europa.eu/digital-single-market/en/news/digital-europe-programme-proposed-eu92-billion-funding-2021-

2027
19 IDC’s Western Europe AI/Cognitive Solutions Survey, June 2018
20 ATI General findings (D3.4), Section 5, May 2020

July 2020

13
AT WATCH- European Commission

data. European enterprises, especially SMEs, are reluctant to share data. Insufficient data
sharing and re-use is a major barrier to the implementation of advanced AI in the supply chains.
Improving the availability of data and data flows across Europe is one of the strategic objectives
of the European Data Strategy21.
• An immature legislative framework. Uncertainty about data governance and the
management of AI-related liability, safety, security risks is slowing down adoption of AI
solutions. Again, this challenge is recognised and addressed by White Paper on AI.
• A limited access to AI data and infrastructure: The lack of AI infrastructures and access to
datasets across borders can represent a bottleneck for scaling up AI deployments, particularly
for innovative entrepreneurs, SMEs and start-ups. Academics and innovators need good access
to world class innovation infrastructure including access to data and resources such as High-
Performance Computing (HPC) and test environments22. Removing this barrier would be
particularly important, since the number of AI start-ups has been growing in the recent period
in Europe23.
• High Complexity of AI in Industry and Public sector: Implementing AI, data and Robotics
relies on incorporating the domain knowledge of underlying processes. Handling these
challenges requires combining domain specific process knowledge with AI based knowledge.24
• A lack of social trust. Public opinion perceives the risks to social cohesion, democracy and
human rights from AI innovation even more than the potential for social good.25 There are also
risks for enterprise reputation and financial welfare. European policies aim at building a
trustworthy AI, but this is not sufficient: enterprises must be aware of these threats and be
proactive in building trust for their customers and markets.
• The presence of several technological barriers: There is considerable complexity and cost in
creating AI systems with the ability to collect, process and analyse large quantities of data in
order to make robust and trustworthy decisions and implement autonomy. The main technology
challenges are presented for example in the BDVA-EU Robotics research agenda26. Enterprises
cannot simply wait for researchers to solve the technology challenges but should work with
them to identify and solve them, particularly in standardisation and interoperability.

Following the COVID-19 outbreak, many industries such as Transportation/logistics, Manufacturing and
Retail will be forced to revise their technology investments downwards. On the other hand, AI is a
technology that can play a significant role in helping businesses and societies to deal with large scale
disruption caused by quarantines and lockdowns. Of all industries, the Government and the Healthcare
sectors, will experience an acceleration of AI investments. For example, hospitals are looking at AI to
speed up COVID-19 diagnosis and testing and to provide automated remote consultations to patients in
self-isolation through chatbots. At the same time, governments are investing in AI to support track and
tracing applications and monitor compliance with social distancing (see also Section 2.3.3 for more
information).

21 European Data Strategy, chapter 4 “Problems”

22 BDVA and EU Robotics, SRIDA, 2019
23 ATI General Findings report (D.3.4), Section 5, May 2020
24 BDVA and EU Robotics, Strategic Research, Innovation and Development Agenda, September 2019
25 Renda, A. (2019). Artificial intelligence: Ethics, governance and policy challenges. CEPS Task Force Report.
26BDVA and EU Robotics, Strategic Research, Innovation and Development Agenda, September 2019

July 2020

14
AT WATCH- European Commission

2.2 Use cases and business opportunities

2.2.1 AI value proposition
The multi-faceted set of technologies falling under the name of AI can be applied to multiple business
processes, thus generating a wide variety of business opportunities. A good overview of the AI-value
proposition can be found in the research and innovation agenda developed jointly by the BDVA (Big
Data Value Association) and EU Robotics research communities for the forthcoming AI Public-Private
Partnership AI27. They include the following:
• By weaving AI into the design, manufacturing, production and deployment processes,
productivity can be raised.
• By using AI to increase autonomy, higher operational flexibility can be achieved.
• By using AI to improve the usability of products and services (e.g. by allowing greater
variations in the human-machine interaction), the user value can be increased and new
customer segments addressed, therefore creating new markets.
• By using AI for supporting complex decision-making processes in dynamic
environments, people can get help in situations of rising complexity (e.g. technical
complexity, increasing volatility in markets).
• By using AI to improve general public services and for social good, social welfare can be
improved substantially.

As the AI landscape matures, application-specific products will emerge and improved commercial
development tools will be made available to industries with lower high technology awareness and higher
barriers to adoption. This is already taking place in the European AI market, driving impressive spending
growth even in those sectors that are normally resistant to new technologies adoption.
2.2.2 Leading AI Use Cases
In this report a use case is defined as a discrete funding effort supporting the implementation of a
technology to achieve a business or societal objective, in a specific business process or application area.
The identification of the main use cases of a new technology helps to pinpoint its market and business
value and to track users’ choices and priorities. Successful use cases demonstrate the viability and
positive return on investment of new technology solutions in concrete application environments, rather
than pilots or trials. Particularly in the digital environment, in fact, there is often a “hype” factor around
innovative technologies, which tends to overstate their maturity and potential business benefits and
may mislead interested users or developers. In the case of a technology such as AI, with a very broad
range of potential applications, but also complex requirements in terms of actual implementation, the
analysis of emerging use cases is especially useful. In this paragraph we provide an overview of the
leading and emerging AI use cases and the related business opportunities for both the technology
providers and the business users.
The way Artificial Intelligence is applied in real-life situations, and within the different industry sectors,
encompasses a mix of horizontal applications and industry-specific scenarios. As with other
advanced technologies, the way AI brings forward the process of digital transformation and industry
modernisation is becoming more and more demand-driven, where industry users tend to direct their
efforts towards a better understanding of the business and operational opportunities offered by the
technologies rather than focusing on the technologies per se.
Figure 4 below provides a snapshot of the top 10 AI use cases in Europe, in terms of value growth
(forecast spending increase) in the period 2019-2023. The size of the bubbles corresponds to the users’
spending for each use case in 2019. Even if the growth rate forecast for 2020 on 2019 should now be
considered with caution, because of the 2020 recession caused by COVID-19, digital markets are
expected to rebound fast from 2021 onwards, as investments are made to fuel the recovery.
The top AI use cases are a mix of horizontal and vertical-specific. For example, the use case Automated
Customer service is widespread across the customer-facing industries, from banking to retail to
telecommunication or even utilities, but it does extend to manufacturing too. As a result, it shows the
highest spending value in 2019 (€413 m, Table 1) and is still expected to grow in the next years. The
Asset/Fleet management use case is specific to the Transport sector (so it has a smaller potential
market), but it provides valuable business benefits and shows the highest expected growth rate to 2023

27 BDVA and EU Robotics, SRIDA, 2019

July 2020

15
AT WATCH- European Commission

(47% CAGR, Compound Annual Growth Rate, Table 1). Different use cases may also share common
business or societal outcomes objectives, such as higher internal automation or enhanced customer
or user experience. These outcomes should be the reference and starting point for those European
organisations looking with interest at the AI landscape in order to capture new business opportunities.
Figure 4: Top 10 AI Use Cases in Europe by value growth (spending growth, € m)

Source: IDC Worldwide Semi-annual Artificial Intelligence Systems Spending Guide, March 2020
Legend: The bubble size represents the spending value by use case in 2019, CAGR: Compound Annual Growth Rate.

Table 1: Top 10 AI Use cases Spending (€ m) (Ranking by Value)

2019 Spending
Top 10 AI Use cases 2019-2023 CAGR 2020/2019 Growth Rate
(€ m)
Automated Customer Service 35% 47% 413

Defense & Government

35% 40% 298
Intelligence

Diagnosis/Treatment Systems 37% 49% 271

Automated Claims Processing 40% 41% 117

Freight Management 41% 44% 111

Advanced Digital Simulation 42% 54% 100

Asset/Fleet Management 47% 51% 98

Shopping Advisory &

42% 51% 90
Recommendations

Automated Content Solutions 38% 40% 75

Smart Networking 35% 60% 35

Source: IDC Worldwide Semi-annual Artificial Intelligence Systems Spending Guide, March 2020. Further information
about the leading use cases is presented below.

July 2020

16
AT WATCH- European Commission

Automated Customer Service

The use of AI in Customer Experience scenarios is one of the most mature applications, enabling faster
and automated customer support via online chatbots and customer enquiries triage tools, improving
productivity and customer satisfaction. AI solutions applied to automated customer services help
understand customer needs and problems and reduce the time and resources spent in achieving
customer issue resolution: they can be adopted in most industries, from manufacturing to healthcare,
but are particularly relevant in consumer-facing industries such as finance and retail. Even though this
use case already generates a high level of spending, it also has a high forecast growth thanks to a
significant potential for further uptake by new users. Moreover, the use of AI in customer interaction is
fast becoming a competitive imperative. This use case is also a valuable business opportunity for
potential suppliers because its implementation often requires personalisation and adaptation services
on top of standard solutions.
Defence and Government Intelligence
Governments around the world are embedding AI functions and applications in their advanced
intelligence practices, to increase the level of automation of the identification and management of public
risks. In this context, AI solutions deliver critical data, information, and intelligence that help national
security agencies identify, monitor and respond to personnel, assets and infrastructure threats.
Diagnosis/ Treatment Systems
AI systems allow to extract insights from the intersection of diverse data sets, including medical records,
lab tests, clinical studies, medical images, and other sources, to assist in diagnosis and provide
personalised treatment at the individual patient level. Hospitals were already implementing this use
case before the COVID-19 pandemic, which has highlighted its usefulness and is likely to accelerate its
adoption. AI is transforming patient care models through intelligent diagnosis assistance and
personalised treatments delivery.
Automated claims processing
The insurance industry is leveraging AI in the automated processing of the data generated by claims to
increase the productivity and effectiveness of their investigation and adjudication. This is a typical case
of data-driven, intelligent decision support innovation which is already generating substantial value for
the industry. The insurance industry is very fragmented and the potential of growth in the next years
with further uptake of this use case is quite high.
Freight Management
The implementation of AI in freight management logistics for all transportation modes (air, rail, land,
and water) combined with supply chain logistics leads to intelligent monitoring and end-to-end visibility,
enabling the optimisation of truck loads, container management, spare parts planning, dock availability,
and customer experience management. In addition, this use cases helps suppliers to gain intelligent
insights on the type of product, place, and time of delivery to guarantee best pricing.
Asset & Fleet Tracking in Transportation
This use case helps to increase visibility of assets and real-time operational intelligence in the transport
and aviation industries improving profitability and productivity. With these tools, it is possible to
implement remote management or resources, allowing to achieve route optimisation; provide actionable
responses to vehicle condition (remote diagnostics); monitor driver behaviour (tracking of idle or
stopped time).
Shopping Advisory & Recommendations
Leading companies such as Amazon or Netflix have built their fortune on profiling customer tastes and
offering personalised advice. The news is that thanks to maturing AI many more retail enterprises can
take advantage of this technology to enhance marketing and advertising practices and to combine
physical and virtual commerce.
Advanced Digital Simulation
This use case is applied in Manufacturing to improve product design and production through a
continuously learning system (deep learning) which can be queried automatically or by voice. The use

July 2020

17
AT WATCH- European Commission

of deep learning and AI results in a virtual model that is essentially the intelligent counterpart to an
actual, physical object (digital twin): in this way engineers can test the simulated product, anticipate
problems, learn about weaknesses. This provides valuable productivity and efficiency benefits, but also
customer satisfaction by allowing to improve the quality of new products.
Automated Content Solutions
This use case leverages AI for the content workflow process. Examples include, but are not limited to,
content marketing, standard research procedures, research reports, white papers, product data sheets,
and journalism (news articles, stories produced automatically by AI rather than human reporters or
authors). This use case is still in the emerging phase but shows high growth potential. Smart
Networking
The telecommunication industry is exploiting AI to expand network management and real-time analytics
applications. In this specific context, AI solutions provide design and implementation of rules and
parameters governing the routing of inbound calling together with an expanded use of real-time
analytics to make the network smarter and more efficient.
In conclusion, AI plays a key accelerator role behind all these established and emerging use cases.
However, it is the combination of AI with other advanced technologies that makes these digital
use cases possible. This is the reason why it becomes so crucial for the European industry to chart
out a forward-looking digital roadmap across different maturity time horizons; a roadmap that, starting
from business and societal outcomes, identifies the leading enabling use cases, while pinpointing at the
same time the technologies required to successfully support the use case implementation.

2.3 Social and Sustainability Impacts

2.3.1 Introduction
The influence of AI-driven predictions, recommendations and decisions is gradually affecting all
dimensions of life. The awareness of the potential positive impacts on social welfare and the economy
is counterbalanced by the emergence of new ethical and fairness risks. Fortunately, the same AI
technologies that generate ethical risks can also be used to fight them and create social good. A case
in point is the spread of fake news, such as “deep-fakes” (realistic, hard to detect and surprisingly easy-
to-create facsimiles of real people28, used for example to spread disinformation in political elections29)
which can be revealed with AI tools. This is another motivation to promote a human-centered and
trustworthy AI.
The achievement of social goals can be complementary or alternative to business goals. We can identify
three main categories of stakeholders engaged in the achievement of social goals, in different and not
mutually exclusive ways:
• Public or private organisations providing general public services, including government, social
services, education, healthcare, public transport and other utilities. In these sectors,
organisations pursue public value in addition to efficiency and effectiveness, to improve social
welfare and quality of life. Technology innovation can be used for example to improve
democracy, transparency, social inclusion, equal access to basic services, the respect of human
rights.
• Public or private organisations investing in technology initiatives for social good, for example in
the context of the achievement of the UN 2030 Sustainable Development Goals (SDG). For
example, Google.org launched an open call30 to identify projects using AI for social good,
collected proposals from 119 countries and selected 20 winner initiatives to be funded. Some
private organisations do so in the context of their Corporate Social Responsibility (CSR)
strategies.
• Governmental or non-governmental organisations and other non-profit organisations (part of
the so-called Third sector) who by definition pursue social goals and can use technology to do
so.
In this chapter we examine the emerging application domains and use cases of AI which contribute to
social welfare, quality of life and environmental sustainability, to raise the awareness of stakeholders

28 https://threatpost.com/deepfakes-ai-fighting-cybersecurity-fire/154978/
29 https://www.stern.nyu.edu/experience-stern/faculty-research/disinformation-and-2020-election-how-social-media-
industry-should-prepare
30 https://ai.google/social-good/impact-challenge

July 2020

18
AT WATCH- European Commission

about this category of impacts and encourage their initiatives in this domain. This is structured in two
main topics:
• The use of AI for social good, reviewing innovative initiatives by international organisations as
well as private and public actors;
• The social impacts of AI in the healthcare sector, with a specific focus on the actual and potential
use of AI to manage the COVID-19 pandemic and its consequences.

2.3.2 AI for social good

The research, business and social communities are becoming increasingly aware of the need to support
and promote the use of AI for social good, with a proactive approach complementing private initiative
where necessary. This is recognised by leading AI policies and guidelines. For example, the first priority
of the OECD Principles on Artificial Intelligence31 (the first international standards agreed by
governments adopted in May 2019) is that human-centred AI should contribute to inclusive and
sustainable growth and well-being. The EC’s “White Paper on Artificial Intelligence - A European
approach to excellence and trust”, published for consultation in February 2020, underlines that the use
of AI systems impacts not only individuals and businesses, but also society as a whole and
“can have a significant role in achieving the UN 2030 Sustainable Development Goals (SDG),
in supporting the democratic process and social rights”. The EC expects AI technologies to play
an important role also in the implementation of the European Green Deal, to deal with environmental
challenges.
The range of social impacts covered by AI for social good is quite broad, as shown by the initiatives
presented in the Table 2 below. A study32 identified already in 2018 more than 160 use cases of AI
applied to the UN SDG domains and generally able to improve social welfare. The areas of impact
include:
• Leveraging AI to provide high quality healthcare in regions with few resources, as for example
the Médicins sans Frontières project to use image recognition and a smartphone app to help
medical staff in remote areas to diagnose antibiograms and prescribe the right treatment.
• Prevent hunger, one of the priority SDG goals, as the American University of Beirut is doing
in Lebanon using machine learning applied to weather and agriculture data to help farmers with
irrigation schedules. AI can help agriculture in poor countries to make a quality leap and fight
the ravages of climate change. Several initiatives are helping to translate AI-driven innovation
in the food-agriculture ecosystem from advanced economies to low-income countries.
• Improving general public services such as waste management. For example, the Grinngo
Indonesia Foundation is pioneering the use of AI-driven image recognition for waste collection
supporting informal-sector waste collectors.
• Several AI-driven initiatives address equality and inclusion. Particularly interesting is the case
of Skillab, a Dutch NGO, which uses the natural language processing and automated translation
capabilities of AI to document migrants’ skills and help them to identify job and career
opportunities in the European labour market.
• AI can also help to ensure security and justice for all, for example the Swiss NGO HURIDOCS
is testing the use of AI to help human rights lawyers extract relevant information from
documents, a type of service available until now only to advanced law firms for profit.
• Supporting social welfare and social services: multiple initiatives are leveraging AI natural
language processing and sentiment analysis capabilities to help social services deal with
requests for help, as for example the Crisis Text Line NGO is doing to optimise the allocation of
telephone calls to counsellors and reduce wait times.

31 https://www.oecd.org/going-digital/ai/principles/
32 https://www.mckinsey.com/featured-insights/artificial-intelligence/applying-artificial-intelligence-for-social-good

July 2020

19
AT WATCH- European Commission

Table 2: AI for Social Good – selected initiatives

Name Description Country Domain Source

AI-driven data models to improve the
Nexleaf https://ai.google/social-
management of vaccine cold chain US Healthcare
analytics good/impact-challenge/
systems in low-income regions

Fondation Using image recognition and a

Médicins smartphone app to help medical staff https://ai.google/social-
FR Healthcare
sans in remote areas to identify good/impact-challenge/
Frontières antibiograms and prescribe treatment

European
Emergency
Piloting AI to support emergency https://eena.org/artificial-
Number
medical service operations for patients BE, FR, IT Healthcare intelligence-eena-corti-
Association
with cardiac arrest outside of hospital project/
(EENA) &
Corti

AI-based data model based on Hunger

Wadwhani farmers-sourced iPhone photos of pest prevention/ https://ai.google/social-
India
AI traps to monitor pests and provide Precision good/impact-challenge/
advice on pesticide use Agriculture

Hunger
American Applying ML to weather and remotely
prevention/ https://ai.google/social-
University sensed agricultural data to provide Lebanon
Precision good/impact-challenge/
of Beirut farmers with irrigation schedules
agriculture

Image-recognition tool for informal-

Gringgo
sector waste collectors and Waste https://ai.google/social-
Indonesia Indonesia
independent waste management Management good/impact-challenge/
Foundation
companies from low-income regions

AI app helping refugees to document

skills in native languages and transfer
Equality and https://ai.google/social-
Skillab BV it into European labour market NL
inclusion good/impact-challenge/
categories to identify potential career
paths

AI-powered two-way translated

communication and personalised
Talking coaching content to guide parents’ Equality and https://ai.google/social-
US
Points engagement with teachers and inclusion good/impact-challenge/
children and overcome linguistic and
cultural barriers

Structured DL system to provide

students with automated, immediate
Equality and https://ai.google/social-
Quill.org feedback on their writing to improve US
inclusion good/impact-challenge/
skills, filling skills gap particularly for
low-income students

NLP and ML methods to extract

Security and https://ai.google/social-
HURIDOCS information and relevant facts from CH
Justice good/impact-challenge/
documents for human rights lawyers

https://www.itu.int/en/ITU-
Artificially Intelligent Legal Information Security and
AILIRA Australia T/AI/Pages/ai-
Research Assistant for free Justice
repository.aspx

AI-driven monitoring and clustering

Information https://ai.google/social-
Full Fact tools for fact-checking of news and UK
verification good/impact-challenge/
other information

Crisis Text NPL and AI-driven data analytics to Social https://ai.google/social-

US
Line, Inc. optimise the allocation of text welfare good/impact-challenge/

July 2020

20
AT WATCH- European Commission

Name Description Country Domain Source

messages to counsellors reducing wait
times and improving effectiveness

Source: shown in the table

The protection of the environment is a domain where AI is expected to have a strong positive
social impact, as shown by the initiatives further presented in Table 3 below. As in the cases of
healthcare and agriculture, there is a trend to transfer advanced solutions pioneered in the private
sector, for example in the mining or energy industries, to the public domain for the public good. A key
trend is the use of ML (machine learning) and DL (Deep learning) techniques to interpret and forecast
data from various sources to detect pollution, illegal activities such as logging, and protect the
environment. This is done for example to detect illegal mines (Colombia) or improve the prediction of
landslides (Pennsylvania) or monitor the air quality (Uganda). The combination of mobile technologies
with DL techniques for bio-acoustic monitoring of threats to rainforests (US) is particularly creative.
NGOs in the environmental sector are also experimenting with AI. For example, the Wildlife Conservation
Society and World Wildlife Fund are adopting an AI-driven system (PAWS) to help prevent poaching in
wildlife reserves.
Table 3: AI for the Environment – selected initiatives

Name Description Country Domain Source

ML-based
Colegio
analysis of
Mayor de
satellite images Environment https://ai.google/social-good/impact-
Nuestra Colombia
to detect illegal monitoring challenge/
Señora del
mines polluting
Rosario
the environment

Mobile
technology
combined with
DL for
Rainforest bioacoustics Environment https://ai.google/social-good/impact-
US
Connection monitoring of monitoring challenge/
threats to
rainforests health
such as illegal
logging

AI to analyse
data from low-
cost air sensors
from Taxis and
Makerere Environment https://ai.google/social-good/impact-
other locations to Uganda
University monitoring challenge/
improve air
quality
monitoring and
forecasting

Global, open
source
monitoring
platform using
images
processing of Environment https://ai.google/social-good/impact-
WattTime satellite images US
Protection challenge/
to measure
emissions of
major fossil fuel
power plants
where local
monitoring is too

July 2020

21
AT WATCH- European Commission

Name Description Country Domain Source

expensive or not
possible

An Artificial
Intelligence
system that
predicts poaching
risk levels in
different areas of
a wildlife
preserve and
helps rangers
PAWS
patrol more https://www.seas.harvard.edu/news/20
(Protection
efficiently. PAWS Environment 19/10/outsmarting-poachers
Assistant Africa
will be integrated Protection
for Wildlife
into SMART
Security)
(Spatial
Monitoring and
Reporting Tool),
a system used by
the Wildlife
Conservation
Society and
World Wildlife
Fund.

Convolutional
Neural Network
for Earthquake
detection and
ConvNetQua location Environment https://github.com/tperol/ConvNetQuak
US
ke (ConvNetQuake) Protection e
open source tool.
It is trained on
data from
Oklahoma (USA).

Development and
use of DL tools to
improve
The prediction of Emergency
Pennsylvani landslides and services/ https://ai.google/social-good/impact-
US
a State create warning Environment challenge/
University systems to protection
minimise
damages and
save lives

Source: shown in the table

2.3.3 AI in the Healthcare Sector

The outbreak of COVID-19 has put enormous pressure on many European healthcare systems,
stretching the available resources and support to breaking limits. However, it has also triggered an
unprecedented demand for digital health technology solutions and has revealed successful solutions
supporting healthcare systems, single hospitals and healthcare providers. Innovative digital solutions
can play an integral role in flattening the curve, limiting the spread of the virus, and assisting in the
treatment of infected individuals. Artificial Intelligence solutions have become one of the first lines of
defence in the pandemic.

July 2020

22
AT WATCH- European Commission

Figure 5: Examples of AI applications at different stages of the COVID-19

Source: OECD, “Using Artificial intelligence to help combat COVID-19”33

Figure 5 shows how AI tools and techniques can help policymakers and the medical community
understand the COVID-19 virus and accelerate research on treatments by rapidly analysing large
volumes of research data. AI text and data mining tools can uncover the virus’ history, transmission,
and diagnostics, management measures, and lessons from previous epidemics.
Specific applications of AI are available at two different levels: at the healthcare system level
and at the hospital level.
Large health-care systems are turning to AI to monitor patients and to regulate the flow of visitors as
they attempt to contain the spread of the novel coronavirus. They are scanning faces to check
temperatures and harnessing fitness tracker data, to zero in on individual cases and potential clusters.
They are also using AI to keep tabs on the virus in their own communities. Leveraging multisource data,
advanced analytics and AI technology, platforms have the potential to help public health bodies in
multiple ways:
• Epidemiological surveillance, identifying early signals of the outbreak and tracking
the infection.
As widely reported, many public health authorities are limiting test eligibility to symptomatic patients
and healthcare workers. Furthermore, there is a time delay between the onset of physical symptoms
and, if the infected individual gets a diagnostic test, the receipt of results. Health authorities can be
more effective, in terms of response times, by accelerating research and make more informed data-
driven decisions, including optimising and triaging of scarce resources and slowing the spread. Skilful
initial selection of optimal tools and approaches improves the investigation. Some existing
epidemiological surveillance platforms have shown promising results using AI and machine learning for
tracing infection spread across time and space and for predictive risk mapping to support epidemic
monitoring response. As an example, the European Centre for Disease Prevention and Control uses NLP
and social listening techniques to analyse data from the internet and social networks to monitor disease
outbreaks. As part of the epidemic intelligence project, data specialists and medical epidemiologists are
analysing the internet, the media and social networks in search of new cases.34
Nevertheless, the success of the global effort to use AI techniques to address the COVID-19 pandemic
hinges upon sufficient access to data. Machine Learning (ML) and Deep Learning (DL) require notoriously
large amounts of data and computing power to develop and train new algorithms and neural network
architectures. Several European countries have worked in this direction.

33 https://read.oecd-ilibrary.org/view/?ref=130_130771-3jtyra9uoh&title=Using-artificial-intelligence-to-help-combat-
COVID-19
34 https://www.ecdc.europa.eu/en/publications-data/epidemic-intelligence-information-system-epis

July 2020

23
AT WATCH- European Commission

• Population screening, detecting and forecasting COVID-19 spread patterns.

Big Data and AI can help identify persons at risk and sites with heightened risk of contagion. Chatbots,
dedicated contact centres and other tools have been put in place to screen patients (asking them
questions about their symptoms and travel history), inform individuals about infection risks and hygiene
practice and manage triage for potential new cases. As an example, the Italian digital health platform
company Paginemediche has developed a chatbot based on the guidelines of the Italian Ministry of
Health, which supports phone triaging of potential cases. However, chatbot developers need to
constantly update the algorithm based on official guidelines, which is rather challenging in a fast-paced
evolving scenario.35
• Contact tracing and social distancing deployment, observing population movements
and monitoring compliance to laws and guidelines.
AI-enabled mobile apps for contact tracing have helped accelerating the process of identifying people
at risk of being infected, thanks to the immediate analysis of the contact network and location history
of individuals. The Italian government has called on start-ups and ICT companies to develop apps to
support smart social distancing, contract tracing and quarantine compliance. To this end, healthcare
service provider Centro Medico Sant Agostino, app developer Bending Spoons, digital marketing
company Jakala and location tech company GeoUniq, have developed a geolocation mobile app to trace
movements of newly diagnosed COVID-19 patients with a margin error of 10 m. Patients may enrol on
a voluntary basis and anonymised data will be shared with healthcare authorities and civil protection.36
When turning from the healthcare system as a whole to the hospital level in particular, it is worth noting
that the latter are being hit by an unprecedented demand for services at an unbearable speed, with
limited organisational resources. The whole system, from clinical to administrative, is put under great
pressure to care for infected as well as usual patients, while limiting virus spread and safeguarding
patients, clinical and non-clinical staff. In this scenario, timing and resource management are the most
critical elements and digital technologies, AI in particular, are playing a primary role to win the race and
provide immediate and efficient clinical decision support.
Hospitals are using AI to help and triage patients and identify those patients that most likely to develop
severe symptoms, thus alleviating the clinical load of physicians. These AI systems have learned about
patterns of illness by processing data from thousands of patient records – and while there is not enough
data from COVID-19 patients yet to create entirely new prediction tools, researchers are checking to
see if existing tools can be customised to help with COVID-19. Most clinical applications of AI to the
COVID-19 response have focused on diagnosis based on medical imaging.
A group of European hospitals and research institutions, for instance, have recently come together to
collaborate in a European Multicentre to automate the diagnosis of COVID-19 on computer tomography
(CT) scans and to quantify disease burden in the lungs of infected patients. The developed AI model will
be made freely accessible to all participating hospitals and institutions. The project is supported by the
European Society of Medical Imaging Informatics (EuSoMII). The research collaboration will be
coordinated by the Netherlands Cancer Institute, who are experienced in applying Artificial Intelligence
to medical imaging.37
Under mounting pressure, hospitals organisations need to shift resources in an agile manner and remove
non-essential workloads away from key staff to lighten their burden. For example, the strain on intensive
care departments can be somehow mitigated by providing clinicians with AI-enabled decision support
tools that can help predict survival rates for patients, speed up diagnosis, monitor patients' vitals
evolution and predict possible complications. AI-based imaging analytics solutions can help in reading
large volumes of medical images. These solutions can help spot lung lesions characteristic to COVID-
19, helping identify high-risk patients and expedite early identification of cases for further confirmation.

35 https://digitalhealthitalia.com/emergenza-coronavirus-parte-la-campagna-tivideovisito-e-in-soccorso-dei-medici-arriva-
la-telemedicina-con-paginemediche-it/
36http://www.salute.gov.it/portale/nuovocoronavirus/dettaglioNotizieNuovoCoronavirus.jsp?lingua=italiano&menu=notizie

&p=dalministero&id=4513
37 https://imagingcovid19ai.eu/#the-project

July 2020

24
AT WATCH- European Commission

2.4 Policy, Regulatory and Ethical Implications

2.4.1 Recent Policy Developments: An Overview
Artificial Intelligence (AI) is poised to span every aspect of our daily lives. According to a recent research
carried out by CEPS, AI has “an enormous potential, [but is] in need of a direction”. Indeed, beyond its
outstanding promise, many commentators also argue that AI, if badly governed, can represent an
existential risk for our society ; whereas others observed that AI can make catastrophic events such as
a nuclear war more likely. In this respect, the AI disruptive potential has not only drawn the attention
of businesses, but also of national and international policymakers and regulators. In the last two years,
the growing attention paid to this topic has resulted in various policy initiatives, both at European and
international level.
The EU has an important role to fulfil in providing consistency among the relevant regulatory
frameworks. In June 2018, the High-Level Expert Group on Artificial Intelligence (AI HLEG) was created
to support the implementation of the European Strategy on Artificial Intelligence. This includes the
elaboration of recommendations on future-related policy development and on ethical, legal and societal
issues related to AI, including socio-economic challenges.38 During the first year of activities, the AI
HLEG focused on two major initiatives:

1. The Ethics Guidelines for Trustworthy Artificial Intelligence (AI) (April 2019)
2. Policy and investment recommendations for trustworthy Artificial Intelligence (June 2019)

The guidelines propose seven key requirements that AI systems should meet: Human agency and
oversight, Technical robustness and safety, Privacy and Data governance, Transparency, Diversity, non-
discrimination and fairness, Societal and environmental well-being and Accountability. A practical
implementation guide is put forward for every key requirement. The second deliverable includes 33
practical recommendations on how to empower and protect humans and society in the AI era, while
creating multi-stakeholders Alliances, which can enact a tailored approach to exploit new technological
opportunities in the Single European Market.39
The European approach was fundamentally endorsed by the OECD, which, in May 2019, adopted
principles40 on Artificial Intelligence inspired by EU guidelines. According to the OECD 41, AI
implementations should be informed by:

1. inclusive growth, sustainability and well-being and they should benefit people and
ultimately the whole planet;
2. the rule of law, democratic values and human rights and they should ensure a fair and just
society;
3. transparency and social responsibility so that users understand AI-based outcomes and are
empowered to possibly challenged them;
4. security and safety so that potential risks can be continuously assessed and managed;
5. accountability for individuals and organisations deploying AI in line with the above principles.

In a complementary move to the five principles, the OECD launched the AI Policy Observatory42 earlier
this year. In this context, the OECD will collaborate with the EU by firstly focusing on building a database
of national AI strategies and policies. The next phase will focus on making EC publications and reports
available on the OECD AI Policy Observatory, sharing data more extensively and collaborating more
closely on the design of improved methodologies for data collection.
In February 2020, the Commission took a fresh look into AI and published the White Paper “A
European approach to excellence and trust”43. The White Paper presented policy options for public
consultation and a “Commission Report on safety and liability implications of AI, the Internet of Things

38 Retrieved at: https://ec.europa.eu/digital-single-market/en/high-level-expert-group-artificial-intelligence

39 More information at: https://ec.europa.eu/digital-single-market/en/news/policy-and-investment-recommendations-
trustworthy-artificial-intelligence
40 https://www.oecd.org/going-digital/ai/principles/
41 ibidem
42 https://www.oecd.ai/
43 See also https://ec.europa.eu/info/publications/white-paper-artificial-intelligence-european-approach-excellence-and-

trust_en

July 2020

25
AT WATCH- European Commission

and Robotics”, providing background information on the type of risks that the White Paper proposes to
deal with (to be followed by legislative proposals in the 4th quarter of 2020).
Figure 6: White Paper on AI- takeaways

Source: White paper “On Artificial Intelligence - A European approach to excellence and trust” Brussels, 19.2.2020
COM(2020) 65 final

The White Paper is organised around two main goals, the creation of an “ecosystem of excellence” to
build European capabilities for AI and an “ecosystem of trust”, to make sure that the legislative system
is updated to deal with new challenges, based on the choice to develop a risk-based approach where
only technologies and solutions with high potential risks for human rights and social welfare are
managed through regulation or other tools. This approach will be finalised and completed based on the
feedback from public consultation.
With this initiative, the Commission encourages governments, businesses and stakeholders to pursue a
uniform approach to AI across Europe, in order to avoid risky barriers in its single market. Moreover,
whenever the AI was deployed in a high-risk sector, the White Paper states that any future legislation
should “specifically and exhaustively” list such sectors.
The Commission complemented the White Paper on AI with a new and comprehensive data strategy,
aiming to make more data sets available for business and government. The strategy aims to promote
AI development, creating “a single market for data will allow it to flow freely within the EU and across
sectors for the benefit of businesses, researchers and public administrations.” 44 The new European Data
strategy outlines the ambition for Europe to become a leading role model for a society empowered by
data to make better decisions in business and the public sector and a global leader in the data-agile
economy.

2.4.2 AI and the need of trustworthiness

The disruptive potential that AI is likely to bring to our society has focused primarily on the ethical and
trustworthiness-related challenges posed by the actual implementation of this technology. While this
threatening narrative should not overshadow the positive effects that AI will exert on our daily lives, it
is important to map possible risks. In fact, Artificial Intelligence is often perceived as a black box
technology, with a lingering fear of unintended negative consequences including the ones that are not
yet known or experienced. Organisations are confronted with significant AI/machine learning (ML)
business risks and potential negative business impact, if the risks are not mitigated appropriately (see
Figure 7 below).

44 https://ec.europa.eu/info/strategy/priorities-2019-2024/europe-fit-digital-age/european-data-strategy

July 2020

26
AT WATCH- European Commission

Figure 7: AI/ML Business Risks and Potential Negative Impact

Source: IDC, 2020

Within this context, trustworthiness is developing as a dominant prerequisite for AI, and relevant
stakeholders must take a proactive stance. The number of firms reporting AI as a risk factor in their
annual shareholder reports (filed to the Securities and Exchange Commission in 2018) has more than
doubled in 2018 according to a Wall Street Journal article.45
A survey conducted in May 201946 found that, while automation, business agility and customer
satisfaction are the primary drivers for AI initiatives, the cost of the solution, the lack of adequate quality
data, the scarcity of data science talent, fear of bias and lack of trust have held organisations back from
implementing AI.
Against this background, an organisation that focuses on being ethical in its AI systems gains
competitive advantage in the long run47. Therefore, there is an intangible value for organisations which
build brand reputation, credibility and trustworthiness in their AI systems. To thrive and reap the
benefits of AI, organisations need to embrace building trusted and ethical AI today and adhere to a
governance framework.
In April 2019, the European Commission presented the “Ethics Guidelines for Trustworthy Artificial
Intelligence” (see Section 2.4.1)48. According to the Guidelines, trustworthy AI should be lawful -
respecting all applicable laws and regulations, ethical - respecting ethical principles and values and
robust - both from a technical perspective and considering its social environment.
Trust, bias and ethics considerations are gaining importance. According to a survey, ~50% of
organisations have a formalised framework to encourage considerations and ~25% have senior
management position established to ensure adherence 49. Diversity of data science teams is growing.
Figure 8 shows the various components of governance framework for trustworthy AI: ethics, elements
of trust, validation/monitoring/analysis and accountability.

45 “More Companies Flag a New Risk: Artificial Intelligence” https://www.wsj.com/articles/more-companies-flag-a-new-risk-

artificial-intelligence-11547035202
46 IDC's AI Global Survey, retrieved at https://www.idc.com/getdoc.jsp?containerId=EUR145364419
47 More evidence can be found at: Floridi, Luciano University of Oxford “The Virtuous Circle of Trusted AI: Turning Ethical

and Transparent AI Into a Competitive Advantage” https://www.capgemini.com/research/the-virtuous-circle-of-trusted-ai-

turning-ethical-and-transparent-ai-into-a-competitive-advantage-luciano-floridi/, https://www.capgemini.com/news/ethics-
in-ai/, Ethical Artificial Intelligence Becomes A Supreme Competitive Advantage
https://www.forbes.com/sites/joemckendrick/2019/07/07/ethical-artificial-intelligence-becomes-a-supreme-competitive-
advantage/#45ed7daa1a8f
48 https://ec.europa.eu/digital-single-market/en/news/ethics-guidelines-trustworthy-ai
49 IDC's AI Global Survey, retrieved at https://www.idc.com/getdoc.jsp?containerId=EUR145364419

July 2020

27
AT WATCH- European Commission

Figure 8: AI Governance Framework for Trustworthy and Ethical AI

Source: IDC, 2020

Fairness is the fundamental aspect of AI that is directly related to discrimination as Machine Learning
could be considered, by its very nature, a form of statistical discrimination. However, discrimination
becomes objectionable when it places certain privileged groups at systematic advantage and certain
unprivileged groups at systematic disadvantage. AI fairness is therefore associated with the
minimisation of bias in AI agents. At the same time, AI systems tend to be opaque and unable to explain
how or why certain results are presented. Understanding how AI models arrive to specific decisions
(explainability) is a key principle of trusted AI. What is more, AI systems should be robust, thus not
vulnerable to tampering or compromising the data they are trained on (robustness), as well as safe to
use (safety) and resilient to all sorts of attacks, including many based on adversarial AI methods
(security). Furthermore, as AI systems are constantly evolving, they should include details on their
development, deployment and maintenance so that they can be audited throughout their life cycle
(lineage). Finally, when discussing trustworthiness as a prerequisite for AI, an element of disclosure
and transparency cannot be ignored. Transparency can be improved, for example, by establishing
factsheets for all the applied AI models, so to help answer basic questions about the dataset models
and include bias mitigation mechanisms.

2.5 Conclusions
Advanced Technologies lie at the very heart of Europe’s current process of digital transformation.
Indeed, these technologies enable process, product, and service innovation throughout the economy,
thus fostering and supporting industrial modernisation across a wide range of organisations and industry
sectors.
While some of these technologies are clearly horizontal in nature, our analysis has emphasised how
other advanced technologies tend to play a more niche and industry-specific role. In this respect the
whole spectrum of the European industry appears to be positively affected by the digitisation brought
about by advanced technologies. Nevertheless, some sectors are taking the lead in the process.
Industries like Manufacturing, Telecommunications & Media and Finance are clearly benefiting from the
uptake of technologies such as Robotics, Augmented Reality/Virtual Reality (AR/VR), 3D printing and
Advanced Materials (especially in manufacturing), or from AI, Big Data & Analytics and Blockchain (in
the finance sector) or again from Cloud Computing, Security. 5G and Photonics (as in the Telco & Media
space). Other industries, as in the case of Transport & Logistics or Agriculture, are showing a slower
pace of advanced technologies’ uptake with technologies such as IoT, AI and Big Data & Analytics
making their way into industry-specific applications in these fields but at a reduced speed.
Taking Artificial Intelligence (AI) as one the backbones of digital transformation and industrial
modernisation, we observed that the current availability of huge datasets, coupled with recent

July 2020

28
AT WATCH- European Commission

technology advances in Big Data, Internet of Things (IoT) and fast connectivity capabilities, are all
paving the way to new and more disruptive applications of AI systems and services. AI’s potential for
future growth and economic impacts is even more significant given that AI is a generic-purpose
technology with the ability to be adopted in different ways by all industries and all social actors. Not
surprisingly, we found that the speed of evolution of the AI market is impressive - the European Union
market alone is expected to grow faster than the global market, representing a share of 23% of total
by 2023 (versus 18% in 2018). Sectors such as Finance, Telecom/media and Utilities/Oil & Gas are
currently leading the way in terms of AI adoption and will continue to do so over the next couple of
years.
Based on the above, we have also investigated how AI is currently being applied in real-life situations.
We found that the top AI use cases are a mix of horizontal and vertical-specific applications with
customer-centric cases (such as Automated Customer Service, for example) being widespread across a
range of customer-facing industries such as banking, retail and telecommunications, and other use
cases (AI in asset or fleet management, for instance) being specific to an individual industry – the
transport sector in this case. We also observed that some use cases have business or societal objectives
in common, even if applied in substantially different sectors and situations. These use cases offer fertile
ground for further applications and extension of AI usage. Businesses and policymakers that are
interested in new AI-related business opportunities and applications are advised to monitor the
deployment of these use cases very closely. Business opportunities, however, are only one side of the
AI picture. As the influence of AI-driven predictions, recommendations and decisions is gradually
affecting all dimensions of life, AI application can exert positive impacts on social welfare and actively
contribute to social good. Our analysis has identified a very broad spectrum of AI applications for social
good, which could range from high-quality healthcare in regions with few resources, to hunger
prevention using machine learning to help farmers with irrigation schedules; from improving general
public services such as waste management to supporting social welfare services by using AI-based
natural language processing to better identify request of help and support social services in better
allocating resources. A case in point of AI for social good, is its application in the healthcare sector.
Here, our research has highlighted how specific AI applications are already at work to support healthcare
systems, and hospitals in particular, to deal with the unprecedented challenges posed by the COVID-19
pandemic. Our research has found that AI is proving to be of paramount importance to improve
epidemiologic surveillance through early identification of the virus outbreak and the tracking of the
infection but is also contributing to better population screening through enhanced detecting and
forecasting spread patterns. In this current phase of loosening of lockdown measures, AI is further
helping authorities to trace contacts and enforce physical distancing, observing population movements
and monitoring compliance to laws and guidelines.
Finally, we have focused on the ethical and trustworthiness-related challenges associated to the actual
implementation of AI. In this context we have emphasised the importance of the European Commission
guidelines for the deployment and application of a trustworthy Artificial Intelligence50 and proposed a
governance framework for trustworthy and ethical AI, which would serve as valuable guidance for
European businesses, researchers and policy-makers in the near future.

50“Ethics Guidelines for Trustworthy AI: High-Level Expert Group on Artificial Intelligence”, 18 April 2020
(https://ec.europa.eu/digital-single-market/en/news/ethics-guidelines-trustworthy-ai)

July 2020

29
AT WATCH- European Commission

Appendix A: Bibliography

Allocato, Adriana and Piai, Silvia. (2020) In Pursuit of a Competitive Position: The Dynamic EHR
Landscape in Europe. Retrieved at https://www.idc.com/getdoc.jsp?containerId=EUR146159620
Allocato, A. Et al. (2020) Health IT in the Fight Against COVID-19: Enabling Resilience, Accelerating
Digital Transformation. Retrieved at https://www.idc.com/getdoc.jsp?containerId=EUR146158020
Cheng, Z. (2020). Huiying Medical: Helping Combat COVID-19 with AI Technology
https://www.intel.com/content/www/us/en/artificial-intelligence/posts/huiying-medical-
COVID19.html
Floridi, Luciano University of Oxford “The Virtuous Circle of Trusted AI: Turning Ethical and Transparent
AI Into a Competitive Advantage” https://www.capgemini.com/research/the-virtuous-circle-of-
trusted-ai-turning-ethical-and-transparent-ai-into-a-competitive-advantage-luciano-floridi/
Geist, Edward and Andrew J. Lohn, (2018), “How Might Artificial Intelligence Affect the Risk of Nuclear
War?”, Santa Monica, CA: RAND Corporation. Retrieved at
https://www.rand.org/pubs/perspectives/PE296.html
Renda, A. (2019). Artificial intelligence: Ethics, governance and policy challenges. CEPS Task Force
Report https://www.ceps.eu/ceps-publications/artificial-intelligence-ethics-governance-and-policy-
challenges/
Vernon, Jack and Spinoni, Erica. (2019). AI in Europe: Key Findings from IDC's 2019 Global AI Survey.
Retrieved at https://www.idc.com/getdoc.jsp?containerId=EUR145364419
Internet sources:
BDVA and EU Robotics, SRIDA-Strategic Research, Innovation and Deployment Agenda, September
2019 http://www.bdva.eu/node/1359
BDVA’s response to the European Commission’s Whitepaper on Artificial Intelligence “A European
approach to excellence and trust” May 2020 - http://www.bdva.eu/sites/default/files/
BDVA%27s%20reponse%20to%20the%20European%20AI%20whitepaper%20-
%20May%202020%20-%20ed1.pdf
Dipartimento della Protezione Civile- COVID-19 Italia - Monitoraggio della situazione, consulted in June
2020,
http://opendatadpc.maps.arcgis.com/apps/opsdashboard/index.html#/b0c68bce2cce478eaac82fe38d
4138b1
European Parliament, Del Ponte, L. (CSIL) 2018, “European Artificial Intelligence. (AI) leadership, the
path for an integrated vision”. European Parliament, Policy Department for Economic, Scientific and
Quality of Life Policies, Directorate-General for Internal Policies.
ECDC (European Centre for Disease Prevention and Control) Epidemic Intelligence Information System
(EPIS) https://www.ecdc.europa.eu/en/publications-data/epidemic-intelligence-information-system-
epis
European Commission Communication, “White Paper on Artificial Intelligence - A European approach to
excellence and trust”, Brussels, 19.2.2020, COM(2020) 65 final
https://ec.europa.eu/info/sites/info/files/commission-white-paper-artificial-intelligence-
feb2020_en.pdf
European Commission “Shaping Europe’s Digital Future”, February 2020, Luxembourg: Publications
Office of the European Union, 2020, https://ec.europa.eu/info/strategy/priorities-2019-
2024/europe-fit-digital-age/shaping-europe-digital-future_en
European Commission Communication, “A European strategy for data”, Brussels, 19.2.2020 COM(2020)
66 final https://ec.europa.eu/info/strategy/priorities-2019-2024/europe-fit-digital-age/european-
data-strategy_en

July 2020

30
AT WATCH- European Commission

European Commission Communication, “Artificial Intelligence for Europe”, Brussels, 25.4.2018,

COM(2018) 237 final https://ec.europa.eu/digital-single-market/en/news/communication-artificial-
intelligence-europe
“Ethics Guidelines for Trustworthy AI: High-Level Expert Group on Artificial Intelligence”, 18 April 2020.
Retrived at https://ec.europa.eu/digital-single-market/en/news/ethics-guidelines-trustworthy-ai
European Commission Communication, “Coordinated Plan on Artificial Intelligence”, Brussels, 7.12.2018
COM(2018) 795 final https://ec.europa.eu/knowledge4policy/publication/coordinated-plan-artificial-
intelligence-com2018-795-final_en
European Commission, Digital Europe Programme: a proposed €9.2 Bn of funding for 2021-2027, June
2019 - https://ec.europa.eu/digital-single-market/en/news/digital-europe-programme-proposed-eu92-
billion-funding-2021-2027
FAO (Food and Agriculture Organization) “Harnessing Artificial Intelligence's transformative power to
make food systems more efficient, sustainable and inclusive”, February 2020
http://www.fao.org/news/story/en/item/1263916/icode/
Forbes, “Ethical Artificial Intelligence Becomes A Supreme Competitive Advantage”, 2019
https://www.forbes.com/sites/joemckendrick/2019/07/07/ethical-artificial-intelligence-becomes-a-
supreme-competitive-advantage/#45ed7daa1a8f
Google.org, AI for social good https://ai.google/social-good/impact-challenge
High Level Expert Group on Artificial Intelligence, set up by the European Commission “Ethical Guidelines
for Trustworthy AI” https://ec.europa.eu/futurium/en/ai-alliance-consultation/guidelines#Top
High Level Expert Group on Artificial Intelligence, set up by the European Commission “Trustworthy AI
assessment list” https://ec.europa.eu/futurium/en/ethics-guidelines-trustworthy-ai/register-
piloting-process-0
IDC’s Western Europe AI/Cognitive Solutions Survey, June 2018
IDC Worldwide Artificial Intelligence Spending Guide 2018–2023, 2H19, March 2020
https://www.idc.com/getdoc.jsp?containerId=US45332319
IDC, “Artificial Intelligence and Machine Learning in Cybersecurity: Creating Meaning with the Terms”,
Market Perspective, February 2019 https://www.idc.com/getdoc.jsp?containerId=US44819319
IDC’s Worldwide Artificial Intelligence Spending Guide Taxonomy, 1H19, 2020,
https://www.idc.com/getdoc.jsp?containerId=US46107920
Imaging AI COVID-19, “Automated diagnosis and quantitative analysis of COVID-19 on imaging”
Consulted in June 2020. Retrieved at https://imagingCOVID19ai.eu/#the-project
ITU (International Telecommunications Union) AI for Good Global Summit, 2020 edition, digital
programme https://aiforgood.itu.int/webinar/
Ministero della Salute Italiano “Contact tracing: Arcuri firma ordinanza per app italiana” April 17,2020
http://www.salute.gov.it/portale/nuovocoronavirus/dettaglioNotizieNuovoCoronavirus.jsp?lingua=itali
ano&menu=notizie&p=dalministero&id=4513
Mc Kinsey Global Institute, Notes from the AI Frontier: “Applying AI for social good”, 2018,
https://www.mckinsey.com/featured-insights/artificial-intelligence/applying-artificial-
intelligence-for-social-good
Mc Kinsey Global Institute, Notes from the AI Frontier: “Tackling Europe’s gap in digital and AI”,
February 2019.
https://www.mckinsey.com/~/media/McKinsey/Featured%20Insights/Artificial%20Intelligence/Tacklin
g%20Europes%20gap%20in%20digital%20and%20AI/MGI-Tackling-Europes-gap-in-digital-and-AI-
Feb-2019-vF.ashx
NYU Stern Center for Business and Human Rights, “Disinformation and the 2020 Election: How the
Social Media Industry Should Prepare”, September 3, 2019 https://www.stern.nyu.edu/experience-
stern/faculty-research/disinformation-and-2020-election-how-social-media-industry-should-prepare

July 2020

31
AT WATCH- European Commission

OECD (2019) “Artificial Intelligence in Society”, OECD Publishing, Paris

https://doi.org/10.1787/eedfee77-en.
OECD, “Recommendation of the Council on Artificial Intelligence” (2019) Available at:
https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0449
OECD Principles on AI, June 2019, https://www.oecd.org/going-digital/ai/principles/
OECD (2019), "Scoping the OECD AI principles: Deliberations of the Expert Group on Artificial
Intelligence at the OECD (AIGO)", OECD Digital Economy Papers, No. 291, OECD Publishing, Paris,
https://doi.org/10.1787/d62f618a-en.
OECD.org/coronavirus “Using artificial Intelligence to help combat COVID-19”, 23 April
2020https://read.oecd-ilibrary.org/view/?ref=130_130771-3jtyra9uoh&title=Using-artificial-
intelligence-to-help-combat-COVID-19
Threatpost, “Deepfakes and AI: Fighting Cybersecurity Fire with Fire”, April 21 2020
https://threatpost.com/deepfakes-ai-fighting-cybersecurity-fire/154978/
UN Secretary General, Big Data and Artificial Intelligence: UN Global Pulse
https://www.unglobalpulse.org/
UNESCO, First ever consensus on Artificial Intelligence and Education published by UNESCO, June 2019,
https://en.unesco.org/news/first-ever-consensus-artificial-intelligence-and-education-published-
unesco
Wall Street Journal, “More Companies Flag a New Risk: Artificial Intelligence”, January 2019
https://www.wsj.com/articles/more-companies-flag-a-new-risk-artificial-intelligence-
11547035202
Robert Koch-Institut: COVID-19-Dashboard, consulted in June 2020. Retrieved at
https://experience.arcgis.com/experience/478220a4c454480e823b17327b2bf1d4
Redazione Digital Health Italia “In soccorso dei medici arriva la telemedicina grazie alla startup
paginemediche” March 13, 2020. Retrieved at https://digitalhealthitalia.com/emergenza-coronavirus-
parte-la-campagna-tivideovisito-e-in-soccorso-dei-medici-arriva-la-telemedicina-con-paginemediche-
it/

July 2020

32
AT WATCH- European Commission

Appendix B: Technology definitions

The advanced technologies covered in the ‘Advanced Technologies for Industry’ project include the
following:
Advanced Manufacturing Technology
Advanced Manufacturing Technology encompasses the use of innovative technology to improve products
or processes that drive innovation in manufacturing. It covers two types of technologies: process
technology that is used to produce any of other advanced technologies, and process technology that is
based on Robotics, automation technology or computer-integrated manufacturing. For the former, such
process technology typically relates to production apparatus, equipment and procedures for the
manufacture of specific materials and components. For the latter, process technology includes
measuring, control and testing devices for machines, machine tools and various areas of automated or
IT-based Manufacturing Technology.
Advanced Materials
Advanced Materials lead both to new reduced cost substitutes to existing materials and to new higher
added-value products and services. Advanced Materials offer major improvements in a wide variety of
different fields, e.g. in aerospace, transport, building and health care. They facilitate recycling, lowering
the carbon footprint and energy demand as well as limiting the need for raw materials that are scarce
in Europe.
Artificial Intelligence
Artificial Intelligence is a term used to describe machines performing human-like cognitive functions
(e.g. learning, understanding, reasoning or interacting). It comprises different forms of cognition and
meaning understanding (e.g. speech recognition, natural language processing) and human interaction
(e.g. signal sensing, smart control, simulators). Artificial Intelligence is a heterogenous field in terms of
its technology base. While some aspects like sensors, chips, robots as well as certain applications like
autonomous driving, logistics or medical instruments refer to hardware components, a relevant part of
AI is rooted in algorithms and software.
Augmented/Virtual Reality
Augmented Reality devices look to overlay digital information or objects with a person’s current view of
reality. As such, the user is able to see his/her surroundings while also seeing the AR content - Virtual
Reality devices place end users into a completely new reality, obscuring the view of their existing reality.
Big Data
Big Data is a term describing the continuous increase in data, and the technologies needed to collect,
store, manage, and analyse them. It is a complex and multidimensional phenomenon, impacting people,
processes and technology. From a technology point of view, Big Data encompasses hardware and
software that integrate, organise, manage, analyse, and present data. It is characterised by "four Vs":
volume, velocity, variety and value. Big Data technologies are new generation of technologies and
architectures, designed to economically extract value from very large volumes of a wide variety of data,
by enabling high-velocity capture, discovery, and/or analysis.
Blockchain
Blockchain is a digital, distributed ledger of transactions or records, in which the ledger stores the
information or data and exists across multiple participants in a peer-to-peer network. Distributed ledgers
technology allows new transactions to be added to an existing chain of transactions using a secure,
digital or cryptographic signature. Blockchain protocols aggregate, validate, and relay transactions
within the blockchain network. Blockchain technology allows the data to exist on a network of instances
or "nodes," allowing for copies of the ledger to exist rather than being managed in one centralised
instance.

July 2020

33
AT WATCH- European Commission

Connectivity
Connectivity refers to all those technologies and services that allow end-users to connect to a
communication network. It encompasses an increasing volume of data, wireless and wired protocols
and standards, and combinations within a single use case or location.
Standard connectivity includes Fixed Voice and Mobile Voice telecom services to allow fixed or mobile
voice communications, but also Fixed Data and Mobile Data services to have access and transfer data
via a network.
Advanced connectivity that is in the focus of the ATI project refers to the rise of Internet of Things
scenarios, where connectivity technology boundaries expand beyond wired and cellular (e.g. 4G, 5G,…)
services to Low Power Wide Area Network (LPWAN), Satellite, and Short Range Wireless technologies.
The survey analysis encompasses all these three elements of the Connectivity definition mentioned
above.
Cloud Computing
Cloud Computing includes the delivery of tools and applications like data storage, servers, databases
and software based on a network of remote servers through the Internet. Cloud Computing services
enable users to store files and applications in a virtual place or the cloud and access all the data via the
Internet.
Public Cloud services that have been explored specifically by the ATI survey are available on public
networks and open to a largely unrestricted universe of potential users. Public clouds are designed for
a market, not a single enterprise.
Industrial Biotechnology
Industrial Biotechnology is the application of biotechnology for the industrial processing and production
of chemicals, materials and fuels. It includes the practice of using microorganisms or components of
micro-organisms like enzymes to generate industrially useful products in a more efficient way (e.g. less
energy use, or less by-products), or generate substances and chemical building blocks with specific
capabilities that conventional petrochemical processes cannot provide. There are many examples of
such bio-based products already on the market. The most mature applications are related to enzymes
used in the food, feed and detergents sectors. More recent applications include the production of
biochemicals and biopolymers from agricultural or forest wastes.
Internet of Things (IoT)
The Internet of Things (IoT) refers to the network of smart, interconnected devices and services that
are capable of sensing or even listening to requests. IoT is an aggregation of endpoints that are uniquely
identifiable and that communicate bi-directionally over a network using some form of automated
connectivity. Objects become interconnected, make themselves recognisable, and acquire intelligence
in the sense that they can communicate information about themselves and access information that has
been provided by another source. The Internet of Things relies on networked sensors to remotely
connect, track and manage products, systems and grids. The Industrial Internet of Things (IIoT) – a
subset of the larger Internet of Things – focuses on the specialised requirements of industrial
applications, such as manufacturing, oil and gas, and utilities. IIoT systems connect non-consumer
devices, used by companies, governments and utility providers in their service delivery.
Micro and Nanoelectronics
Micro and Nanoelectronics deal with semiconductor components and highly miniaturised electronic
subsystems and their integration in larger products and systems. They include the fabrication, the
design, the packaging and testing from nano-scale transistors to micro-scale systems integrating
multiple functions on a chip.
Mobility
IT for Mobility
Mobility covers a large number of different technology areas and markets, which does not only
encompass vehicles that take people from point A to point B, but also includes all kinds of technologies

July 2020

34
AT WATCH- European Commission

that make people more mobile (like for example mobile phones etc.). These, however, consist of a large
set of sub-technologies that are hard to capture at the same time. In this project, the patent, trade,
prodcom, investment and skills analysis focus on a sub-section of mobility, which is related to vehicles
only, e.g. satellite navigation and radio-location, which are also the core technologies that are necessary
to make autonomous driving work.
Enterprise Mobility
The survey analysis captures mobility in terms of the workforce. The Enterprise Mobility market is made
up of a conglomeration of mobile solutions and technologies, including hardware, software and services,
empowering a borderless workforce to securely work anywhere, at any time and from any device. It
does not include only the provision of smartphones or tablets to the workforce but also all the tools and
applications for transforming key processes, from internal operations to operations with customers and
suppliers, all the way from the shop floor to the top floor and from the back office to the end customers.
Nanotechnology
Nanotechnology is an umbrella term that covers the design, characterisation, production and application
of structures, devices and systems by controlling shape and size at nanometer scale. Nanotechnology
holds the promise of leading to the development of smart nano and micro devices and systems and to
radical breakthroughs in vital fields such as healthcare, energy, environment and manufacturing.
Photonics
Photonics is a multidisciplinary domain dealing with light, encompassing its generation, detection and
management. Among other things it provides the technological basis for the economic conversion of
sunlight to electricity which is important for the production of renewable energy, and a variety of
electronic components and equipment such as photodiodes, LEDs and lasers.
Robotics
Robotics is technology that encompasses the design, building, implementation, and operation of robots.
Robotics is often organised into three categories: 1) Application specific. This includes Robotics designed
to conduct a specific task or series of tasks for commercial purposes. These robots may be stationary
or mobile but are limited in function as defined by the intended application. 2) Multipurpose.
Multipurpose robots are capable of performing a variety of functions and movements determined by a
user that programs the robot for tasks, movement, range, and other functions and that may change the
effector based on the required task. These robots function autonomously within the parameters of their
programming to conduct tasks for commercial applications and may be fixed, "moveable," or mobile. 3)
Cognitive. Cognitive robots are capable of decision making and reason, which allows them to function
within a complex environment. These robots can learn and make decisions to support optimal function
and performance and are designed for commercial applications. When measuring production and uptake
of Robotics, industrial applications will be taken into account.
Security
Security products are tools designed using a wide variety of technologies to enhance the security of an
organisation's networking infrastructure — including computers, information systems, internet
communications, networks, transactions, personal devices, mainframe, and the cloud — as well as help
provide advanced value-added services and capabilities. Cybersecurity products are utilised to provide
confidentiality, integrity, privacy, and assurance. Through the use of security applications, organisations
are able to provide security management, access control, authentication, malware protection,
encryption, data loss prevention (DLP), intrusion detection and prevention (IDP), vulnerability
assessment (VA), and perimeter defense, among other capabilities.

July 2020

35
AT WATCH- European Commission

Appendix C: Advanced Technology uptake

Figure 9: Advanced Technologies Uptake by European Industries - Question: Which of the following technologies is your organisation using or planning to use?
Technology Finance Gov/Edu Healthcare Manufacturing - discrete Manufacturing - process Professional Services Retail, Wholesale Telecom, Media Transport/Logistics Utilities, Oil, Gas Agriculture
Public Cloud 56% 63% 76% 82% 68% 81% 84% 82% 81% 73% 78%
Big Data and analytics solutions 79% 44% 49% 45% 36% 61% 53% 63% 34% 47% 25%
Mobile solutions 53% 58% 40% 34% 32% 49% 56% 61% 32% 47% 38%
IoT 72% 48% 44% 58% 45% 58% 55% 73% 36% 64% 30%
AI 74% 33% 41% 45% 37% 60% 51% 70% 32% 60% 11%
Robotics 7% 1% 18% 72% 64% 7% 21% 7% 29% 53% 18%
ARVR 21% 38% 36% 10% 5% 10% 26% 48% 3% 13% 0%
Blockchain 62% 3% 1% 1% 1% 3% 4% 7% 1% 0% 0%
Security 88% 79% 78% 79% 73% 73% 75% 79% 68% 81% 51%
Advanced Manufacturing 0% 0% 0% 92% 79% 0% 0% 0% 0% 0% 0%
Advanced Connectivity 58% 63% 48% 37% 37% 54% 42% 82% 34% 63% 37%
Nanotechnologies 3% 0% 33% 37% 9% 3% 1% 6% 1% 19% 0%
Micro and nano electronics 2% 0% 4% 32% 5% 2% 1% 6% 1% 4% 0%
Advanced Materials 2% 0% 19% 24% 27% 2% 1% 3% 9% 13% 0%
Biotechnology 2% 0% 39% 0% 5% 2% 0% 0% 0% 14% 1%
Photonics 2% 1% 19% 24% 14% 1% 1% 12% 0% 19% 19%

Source: Advanced Technologies for Industry Survey, July 2019

Legend: Sum of % of respondents already using or planning to use the technology

July 2020

36 36
AT WATCH- European Commission

About the ‘Advanced Technologies for Industry’ project

The EU’s industrial policy strategy promotes the creation of a competitive European industry. In order
to properly support the implementation of policies and initiatives, a systematic monitoring of
technological trends and reliable, up-to-date data on advanced technologies is needed. To this end, the
Advanced Technologies for Industry (ATI) project has been set up. The project provides policymakers,
industry representatives and academia with:
• Statistical data on the production and use of advanced technologies including enabling
conditions such as skills, investment or entrepreneurship;
• Analytical reports such as on technological trends, sectoral insights and products;
• Analyses of policy measures and policy tools related to the uptake of advanced technologies;
• Analysis of technological trends in competing economies such as in the US, China or Japan;
• Access to technology centres and innovation hubs across EU countries.

You may find more information about the 16 technologies here: https://ati.ec.europa.eu.
The project is undertaken on behalf of the European Commission, Directorate General for Internal
Market, Industry, Entrepreneurship and SMEs and the Executive Agency for Small and Medium-sized
Enterprises (EASME) by IDC, Technopolis Group, Capgemini, Fraunhofer, IDEA Consult and NESTA.

July 2020

37
AT WATCH- European Commission