Scribd
Upload
39 views1 page

Tutorial 11 - AI & Governance 3

The document outlines exercises related to AI ethics and governance, focusing on principles from the European Commission's Recommendations for AI Governance. It discusses issues such as respect for human autonomy, fairness, data lineage, inherent biases, and the risks of feedback loops in AI systems. Additionally, it highlights the importance of technical robustness and safety in AI systems.

Uploaded by

analindatoh.at
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
39 views1 page

Tutorial 11 - AI & Governance 3

The document outlines exercises related to AI ethics and governance, focusing on principles from the European Commission's Recommendations for AI Governance. It discusses issues such as respect for human autonomy, fairness, data lineage, inherent biases, and the risks of feedback loops in AI systems. Additionally, it highlights the importance of technical robustness and safety in AI systems.

Uploaded by

analindatoh.at
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 1

Official (Open)

GRC

Tutorial 11
AI Ethics & Governance 3
Exercises

1. A technology company implements a system for its customers to automate some


of the services. However, the system begins to make decisions on the customers’
behalf without notifying them. Which of the principles of the EC’s
Recommendations for AI Governance does this violate?
Respect for human autonomy
^^the ai did not consult the customers before making
decision nor did it notify them
2. An AI system in a bank is used to determine the credit limit for its customers. It
was realized that the credit limit extended to its customers are consistently
lower for females than males. Which of the principles of the EC’s
Recommendations for AI Governance does this violate?
Fairness

3. A cyber intelligence system with AI capabilities takes, as its learning inputs, cyber
intelligence data from the public domain. Which type of data lineage does the
intelligence data adopt? What is the risk of such a system, with regards to data
accountability? Suggest a mitigation.
the risk of such system would be the quality of the data . Since it is from the public domain , the sources
Forward data lineage
cannor be controlled. Hence malicious actors may intentionally posion the data bu submitting erroneous
data so as to affect the AI systme . A possible mitigation is to impose qlty checks on the data before it is fed
4. When collecting data from real life situations, inherent bias can occur. List and
to the AI system. trends can be monitored to see if there are any anomalies

describe TWO biases.


The two biases are selection bias and measurement biases .
Selection bias is a dataset consisting of data which is not fully representative of the actual env .
MB : a dataset obtainied using a device that gives a certain tendency towards a particular direction
5. What may occur if the data produced by an AI system is feedback into training
itself? reinforcement bias can occur , this leads to increased bias in AI system

It might lead to risk of malfunction in AI systems , such as misclassfication and biaseness

6. Under the requirements of EC’s Ethics Guidelines for Trustworthy AI, technical
robustness and safety means
a) AI systems are reliable and produce accurate results, even in adverse
conditions
b) AI systems use cryptography to keep data confidential
c) AI systems keep private data secret
d) AI systems have traceability features

7. One way of implementing data lineage is


a) Data tagging
b) Data encryption
c) Data hashing
d) Data mining

GRC Page 1

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy