SEEING IS NOT BELIEVING

------Disinformation and AI Deepfake

GROUP 2
22257365 LIU Keru
22257381 LIU Yutong
22253734 Wang Jing
21206449 RUI Xiya
22257780 CHENG Xinya
22223606 Wang Ludan
22257012 Jiang Xinran
OUTLINE
01 Introduction

Coaching activities: how

04 to identify disinfomation
Recent cases and current
02 situation

05 Conclusion
Why deepfake are
03 used& impact of the use
 PART1

1 INTRODUCTION
PART1

WHAT IS DEEPFAKE
Deepfakes are synthetic media where a
person's likeness in an existing image or video
is replaced with someone else's.

Refers to a type of artificial intelligence

technology that creates realistic fake
videos or images.

Powerful tools for generating disinformation.

 PART1
WHAT IS
DISINFORMATION
It involves
Refers to false orchestrated efforts Pose significant
information that is by individuals or challenges to our
deliberately spread groups who use ability to discern
to deceive people strategic deception truth in an
and media increasingly
manipulation digital world ！
tactics to achieve
political, military, or
commercial
objectives
PART1

DISINFORMATION AND
DEEPFAKE
Initially generated for legal entertainment and
education purposes

Used extensively for illegal and unethical

purposes

Many perpetrators of misuse are teenagers

themselves ！
PART 1

OUR PURPOSE Educating and empowering you

to critically evaluate and
navigate media content.

Provide you with the tools

and knowledge to discern and
responsibly engage with
media and information.
2 RECENT CASE
& CURRENT
SITUATION
 part 2
CASES OF DEEPFAKE

POLITICAL PERSONAL IMAGE CRPTO SCAM

DISINFORMATION THEFT

• $25 MILLION
• DEEPFAKE AUDIO
• SOUTH KOREAN FRAUD IN AN
OF JOE BIDEN IN
DEEPFAKE CASE INTERNATIONAL
US ELECTION
COMPANY IN HK
DEEPFAKE AUDIO OF part 2
BIDEN IN US ELECTION
part 2

SOUTH KOREAN DEEPFAKE

CASES

Telegram-based group chats that produce fake

pornographic images of girls and women without
their permission.

South Korean singers and actresses have accounted

for roughly half of the individuals featured in
deepfake pornography online.
part 2

$25 MILLION
FRAUD IN AN
INTERNATIONA
L COMPANY IN
HK
part 2
Deepfake technology has become
a mature application of
THEREFORE disinformation and has widely

... affected our daily lives.

While, the government and

organization’s regulations haven’t
kept pace with the technology.

We should raise awareness of

it and strengthen ourselves
to protect our privacy and
safety.
 3
WHY &
IMAPCT
Making Money - Identity Theft
Fraud

• This employee was tricked into

participating in a video call with what
he thought were other staff members of
the companies
• But in fact these staff members were all
deep fakes
Making Money - Identity Theft
Fraud
• Scammers use deepfake to modify
publicly available video and other
footage to scam people out of their
money
• At least 20 occasions, deepfake AI were
used to mimic the portraits on ID cards
and spoof facial recognition programs

• 30% of companies have experienced

fraud based on deepfake in the past
year.
Making Money - Pornographic
Industry
• The fake images of Taylor
Swift were predominantly
circulating on social media
site X
• The photos, which show the
singer in sexually suggestive
and explicit positions, were
viewed tens of millions of
times
Making Money - Pornographic
Industry

• Earned a large amount of

profits

• Selling deepfake photos and

videos of famous star

• Deepfake sexual photos of

children and adolescents
Influence Politics

Michal Šimečka, leader of the liberal Progressive Slovakia

party
how to rig the election,
partly by buying votes from the with the daily newspaper
country’s marginalized Roma
journalist
minority.

raising beer price

during the election's "silence

period"
• the election being “the first swung by deepfakes”

• introduces challenges to the democratic process --

deprive the public of the accurate information
needed to make informed decisions in elections.

• the deeper concern lies in the emergence of the

"liar's dividend" — a phenomenon where the very
existence of generative AI engenders an
atmosphere of mistrust.
Cyberbullyi
ng
• A court in south-west Spain has sentenced 15
schoolchildren to a year’s probation for creating
and spreading AI-generated images of their
female peers

• obtain manipulated images of by taking

girls’original faces from their social media profiles

• The manipulated photos were then shared on two

WhatsApp groups.
 • Teen cyberbullying

• gain laughter or recognition among

their peers and spread rumour.

• Cyber violence against women

• Targeting female friends, relatives

and acquaintances for sexism or
misogyny
 4 HOW TO
IDENTIFY
DISINFOMATION
coaching activities
1) HOW TO IDENTIFY
DISINFOMATION (FAKE
NEWS) ?
Use this check list:
• source
(credible/original/sponser?)
• headline (outrageous?)
• author (credible/real?)
• supporting source
(accurate?)
• date (new/old?)
• joke (outlandish/funny?)
• personal bias (be critical)
• ask expert (fact check)
Reference:
A guide developed by the Research Center at the CUNY Graduate
School of Journalism: https://library.csi.cuny.edu/c.php?
g=619342&p=4376665#s-lg-box-13405250
FACT CHECK WEBSITE
RECOMMENDATION

Here are the popular fact-

check website with the link:
• PolitiFact.
• FactCheck.org.
• Snopes.
• BBC Verify.
• HKBU fact check.

These websites use

independent fact-checkers
to review the accuracy of
acting and information by
checking reputable media
sources.
 CASE STUDY Let’s check fake news on socia
media!
Two popular posts claimed or implied that people should not get COVID-19 vaccines and
that the risks of vaccines outweigh the benefits ：
 CASE STUDY
Two popular posts claimed or implied that people should not get vaccinated and that the
risks of vaccines outweigh the benefits ：

author: not credible, ordinary

people

Exaggerated headlines

blame
make you
angry

check source they cite

from more credible
news agency, open
them to check the fact
 CASE STUDY The truth is here!
The study confirmed the known side effects of the COVID-19 vaccine, but the post did not
indicate that these side effects were rare, and that the vaccine was very effective in
preventing serious illnesses after infection with COVID-19.

Or go t
o fact c
heck w
ebsite:
 SUMMARY
• Clickbait: uses exaggerated, questionable or misleading
01 Categories of fake news:
headlines to attract readers.
• Propaganda: spreads information to harm an institution,
country, group of people or individual -- typically for political
gain.
• Imposter content: impersonates general news sites
• Biased/slanted news.
• Satire: for parody and entertainment.
• State-sponsored news.

• Be attention to ALL CAPS, or obviously photoshopped

pics.
02 Other tips: • A glut of pop-ups and banner ads? Good sign the story is
pure clickbait.
• Check the domain! Fake sites often add “.co” to trust
brands to steal their luster. (Think: “abcnews.com.co”)
Reference: https://en.wikipedia.org/wiki/List_of_fake_news_websites
https://www.techtarget.com/whatis/feature/10-ways-to-spot- • If you land on an unknown site, check its “About” page.
disinformation-on-social-media • Gut check. If a story makes you angry, it’s probably
https://library.csi.cuny.edu/c.php?g=619342&p=4376665 designed that way.
• If you’re not sure it’s true, don’t share it!
 2) HOW TO IDENTIFY
DEEPFAKE PHOTO OR VIDEO?
Identifying by our eyes:
Observation of Reflections and Shadows.
1) Check the points of light in people’s eyes.
LIGHT AT THE e.g. The picture of Trump:
CENTER
Couldn’t find any light in his eyes (center position)
NO LIGHT FROM Women around him have light (size and color different)
LIGHT LEFT
on their eyes with one fixed light location.

Q ： What is your feeling when you look at the figure in

this picture?
A: Their skin texture and light on their face are too
clear and flareless.

2) Artificial intelligence can also produce discrete （不协调

的） lighting and shadows...AI generated face are
perceived as more like a perfect human.
2) HOW TO IDENTIFY
DEEPFAKE PHOTO OR VIDEO?

Identifying by our eyes:

Hands and ears

AI is currently poor at rendering hands and
ears, mangling their shapes, proportions and
even the number of fingers.

AI-generated images of people become hyper-

real, these inaccuracies create an uncanny
valley.
Sometimes, we can identify deepfake photo or
common AI generated photo by counting how
many ears of eyes of the character have.
2) HOW TO IDENTIFY
DEEPFAKE PHOTO OR
VIDEO?
“The use of fact-checking
tools is a skill that fact-
checking machines can help
you acquire in a short period of
time as a way to improve their
ability to verify the authenticity
of information.” WHERE DOES THE IMAGE SHOW UP?

WHERE DID IT ORIGINATE?

CONTRAST DIFFERENT WEBSITE SEARCH

2) Using some effective app CONSEQUENCES, AND ASK YOURSELF IF
THE IMAGE HAS BEEN ALTERED? IN
or website to help WHAT WAY?
recognize:

1. Google Search engine to

do the reverse Image
2) HOW TO IDENTIFY
DEEPFAKE PHOTO
OR VIDEO?

2. FotoForensics
2) Using some effective app or website to
help recognize:

This tool provides an in-depth analysis of

images on the internet. Identify hidden
pixels, error level analysis and metadata
details → providing evidence for your
judgment of some footage information.
3) MEDIA LITERACY TIPS
Tip 1: Create Responsible Content

01 ACCURACY 04 AUTHENTICITY

02 TRANSPARENCY 05 ENGAGEMENT

03 RESPECTFULNESS
MEDIA LITERACY TIPS
Tip 1: Create Responsible
Content
Scenario: Your school's sports team competes in a
regional tournament and narrowly loses in the final
match, missing out on the championship title by a
small margin.

A. 🏀 Despite falling short in the game, we hold our

heads high. It wasn't our day for the championship,
but we'll learn from this setback and come back
stronger next time! 💪

B. 🔥 Unbelievable loss! We've been outperforming

them in practice, so how did we lose? Could the
other team have cheated? Were the referees bribed
or did they use performance-enhancing drugs? 🤔
MEDIA LITERACY TIPS
Tip 2: Responsible Viewing and
Sharing

01 VERIFY SOURCES 04 RESPECT PRIVACY

02 CRITICAL THINKING 05 PROMOTE POSITIVITY

03 FACT-CHECK
MEDIA LITERACY TIPS
Tip 2: Responsible Viewing and
Sharing
Scenario: The official release of a local natural
disaster, a flooding incident, has been captured in a
video. The video serves as a warning to the public to
stay safe and evacuate immediately.

A. Emergency rescue hotline: XXX-XXX-XXXX. Please

evacuate to designated safe points immediately.
Resource distribution: Food, water, and medical aid
points set up at X location.

B. The flooding is extremely severe, already causing

suffering for a hundred thousand people!

C. How incompetent is the government, taking

people's money but doing nothing in the face of
disaster. There must be corruption involved!
#FloodCrisis #GovernmentFailure
4) HOW TO PROTECT
YOUR PRIVACY
HOW TO PROTECT YOUR PRIVACY
HOW TO
PROTECT YOUR PRIVACY SETTINGS FRIEND REQUESTS

PRIVACY
STRONG PASSWORDS AVOID GEOTAGGING

PERSONAL INFORMATION APP PERMISSIONS

AND………………?
5 CONCLUSION
CONCLUSION

The rise of deep fake technology and the Fostering media literacy
proliferation of disinformation present significant and promoting
challenges to our society responsible viewing and
sharing practices

Better protect ourselves

Individuals must become educated consumers of
and our communities from
media, developing the skills necessary to evaluate
the harmful effects of
information sources and identify signs of deception
disinformation
critically
Referenc
• Stellino, M. (n.d.). 9 tools for verifying images | International Journalists’ Network. Retrieved from https://ijnet.org/en/story/9-tools-verifying-

es
images
• https:/www.bbc.com/future/article/20240311-how-to-spot-a-fake-photohttps:/newscollab.org/2019/02/04/9-tools-to-identify-fake-images-and-
videos/
• https:/zh.gijn.org/stories/%E4%BA%8B%E5%AE%9E%E6%A0%B8%E6%9F%A5%E9%AB%98%E6%89%8B%E4%B8%89%E6%AC%BE%
E5%BC%BA%E5%8A%9B%E5%B7%A5%E5%85%B7-%E8%BE%A8%E5%88%AB%E7%BD%91%E7%BB%9C%E5%9B%BE%E7%89%87
%E7%9C%9F%E5%81%87/
• https://www.mk.co.kr/en/society/11138267
• https://edition.cnn.com/2024/01/25/tech/taylor-swift-ai-generated-images/index.html

• Misinformation and disinformation: Thinking critically about information sources. (2024, October 9). CSI Library CSI Library at CUNY
College of Staten Island Library. https://library.csi.cuny.edu/c.php?g=619342&p=4376665#s-lg-box-13405250
• https://www.bbc.com/news/articles/cpdlpj9zn9go
• https://edition.cnn.com/2024/02/04/asia/deepfake-cfo-scam-hong-kong-intl-hnk/index.
• html https://edition.cnn.com/2024/01/24/politics/deepfake-politician-biden-what-matters
• Diaz Ruiz, Carlos (2023). "Disinformation on digital media platforms: A market-shaping approach". New Media & Society. Online
first: 1–24. doi:10.1177/14614448231207644. S2CID 264816011. This article incorporates text from this source, which is available
under the CC BY 4.0 license
• Yandell, K. (2024, February 29). Study largely confirms known, rare COVID-19 vaccine side effects. FactCheck.org.
https://www.factcheck.org/2024/02/study-largely-confirms-known-rare-covid-19-vaccine-side-effects/
• Jason Gale. (2024, February 19). Largest Covid Vaccine Study Yet Finds Links to Health Conditions. Bloomberg .
https://www.bloomberg.com/news/articles/2024-02-19/largest-covid-vaccine-study-yet-finds-links-to-health-conditions?embedded-
checkout=true
• COVID vaccines linked to slight increases in heart, brain, blood disorders: Study. (2024, February 21). New York Post.
https://nypost.com/2024/02/20/us-news/covid-vaccines-linked-to-slight-increases-in-heart-brain-blood-disorders/
• Amanda Hetler. (2024, October 8). 10 ways to spot disinformation on social media. WhatIs.com.
https://www.techtarget.com/whatis/feature/10-ways-to-spot-disinformation-on-social-media
• https://en.wikipedia.org/wiki/List_of_fake_news_websites
THANKS FOR
LISTENING
Q&A