How Modern Tech is Undermining Truth How Modern Tech is Undermining Truth

and What It Means for the Future

Countries with biggest increases in deepfake-specific fraud

Click to choose a continent and view AI fraud stats

North America
South America
Africa
Europe
Asia

477 Canada
700 Mexico
3000 U.S.A.
828 Brazil
766 Argentina
1000 Algeria
1000 U.A.E.
1200 South Africa
2980 Belgium
1500 Slovakia
1766 Romania
3050 Vietnam
2800 Japan
4500 Philippines

Deepfakes: Trick or Threat?

advantages

disadvantages

AI-generated content

Deepfakes are misused to create non-consensual explicit content, targeting public figures like Taylor Swift and Xochitl Gomez, and anyone sharing photos online

manipulation of words or images

Dangerous precedents include deepfakes of Joe Biden and Slovak politician Michal Simecka used to undermine elections.

Face-swapping actors in movies

One of the most well-known examples of this is ‘Fast 7’ which ‘featured’ Paul Walker—an actor who passed away in 2013 while the movie was in production.

interesting presentations

Deepfake algorithms animate historical figures, making speeches and presentations more engaging than traditional lectures or textbooks.

Dynamic product demonstrations

AI-generated models could showcase products like clothing digitally without expensive photo shoots. Virtual shopping experiences are also possible.

How to spot a deepfake?

Deepfake detection techniques: key Indicators

Facial and Body Movements:

AI struggles with facial expressions and movements, causing an "uncanny valley" effect

Audio Distortions:

Added audio noise can hint at deepfake manipulations meant to conceal alterations.

Unnatural Blinking:

Deepfakes often show unnatural or missing blinking due to AI's difficulty in simulating this behavior.

Shadows and Reflections:

Deepfakes struggle with reflections and shadows. Look for irregularities on surfaces or in the eyes.

Pupil Dilation Issues:

AI struggles to replicate natural dilation, causing unnatural-looking eyes.

Lip-Sync Issues:

Altered audio with video often leads to mismatched lip movements. Watch for misalignment.

Facial and Body Movements:

AI struggles with facial expressions and movements, causing an "uncanny valley" effect

Audio Distortions:

Added audio noise can hint at deepfake manipulations meant to conceal alterations.

Unnatural Blinking:

Deepfakes often show unnatural or missing blinking due to AI's difficulty simulating this behavior.

Shadows and Reflections:

Deepfakes struggle with reflections and shadows. Look for irregularities on surfaces or in eyes.

Pupil Dilation Issues:

AI struggles to replicate natural dilation, causing unnatural-looking eyes.

Lip-Sync Issues:

Altered audio with video often leads to mismatched lip movements. Watch for misalignment.

Your data could be used against you

Identity fraud rates

click here to open

2021

2022

2023

0 0.5% 1.0% 1.5% 2.0%

Top 5 industries with the highest percentage of identity fraud

click here to open

Online Media 4.27%
Professional Services 3.14%
Healthcare 2.41%
Transportation 2.03%
Videogaming 1.82%

Percentage of identity document fraud by country

click here to open

Palau

21.78
Guinea

13.65
Afganistan

10.19
Chad

7.02
Bangladesh

6.11
Pakistan

4.85
Yemen

4.74
Bahrain

4.71
Tanzania

4.61
UK

4.33

0,00% 15,00% 30,00%

How Deepfakes Exploit Emotions: Fear, Urgency, and Trust

Trust in Visual and Audio Content

People tend to trust what they see and hear, especially when it seems authentic or comes from a familiar source. Deepfakes exploit this trust by creating realistic manipulated content, like videos of famous figures, leading viewers to believe the information without verification.

Manipulation Through Fear and Urgency

Deepfake videos can trigger strong emotions like fear or urgency, such as a fake emergency or evacuation call, leading to impulsive reactions based on disinformation without verifying authenticity. Emotions often drive decisions without rational evaluation.

Confirmation Bias

Deepfakes are crafted to align with the audience's existing beliefs, exploiting the tendency to seek confirmation of views. When content matches personal beliefs, people are more likely to accept false information, making deepfakes highly effective at manipulating public opinion.

Fight digital deception with deepfake

Protecting yourself from deepfakes is nearly impossible. Technology is advancing rapidly, but we have a few tips that can still improve your safety online:

Managing Personal Information

Share less personal data online, especially high-quality media. Adjust privacy settings to limit access to trusted contacts.

Secure
Your
Accounts

Use multi-factor authentication (MFA) and strong, unique passwords to secure accounts and protect personal information.

Control Over Personal Information:

Limit the amount of personal data you share online, especially high-quality photos and videos. Use privacy settings to restrict visibility to only trusted contacts.

Responding
to
Deepfakes

If you find deepfake content, report it to the platform and seek help from law enforcement or legal experts.

technological assistance

created by ai 100%

Microsoft Video Authenticator Tool

Microsoft's Video Authenticator uses AI to detect deepfake edits by spotting subtle flaws like fading and grayscale elements invisible tothe human eye.

Attestiv

Attestiv's 2024 update uses ML to detect deepfakes and authenticate media, highlighting potential manipulations with pixel-level heatmaps showing potential manipulations.

created by ai 100%

DeepMedia.AI’s

DeepID detects deepfakes by analyzing pixel-level modifications, image artifacts, pitch, tone, and spectral patterns for audio, and facial expressions and body movements for video.

Sensity AI

Uses a proprietary API to detect deepfakes in media by identifying artifacts and high-frequency signals typical of AI-generated content.

Can you spot a Deepfake?

AI-generated or Real?

Take a look at this image and share whether you think it is generated by AI or not, and how confident you are in this judgment. You have unlimited time to look at the image.

Choose whether the image is fake or real to verify.

Your score is /4

Which means you need to be more cautious on the Internet. Be mindful of deepfake content, as it can be challenging to differentiate between real and fake information. Additionally, always double-check the sources of news and avoid sharing sensitive personal information, as it may increase the risk of misuse or online scams. Stay vigilant and protect your online presence!

Which shows you’re making progress in understanding online risks, but there’s still room to grow. Stay cautious about deepfake content, as it can sometimes be tricky to identify. Make it a habit to verify news sources and think twice before sharing sensitive personal information to avoid potential misuse. You're on your way to becoming more vigilant and secure in your online activities – keep up the effort!

Which means you’re already mindful of some online risks! To continue on this positive path, keep sharpening your awareness on the Internet. Be attentive to deepfake content and always verify sources – you're on the right track to becoming a savvy and secure online user. Great job, and keep up the good work in protecting your online presence!

scroll for more

Fake

This is a synthetic image produced by AI. Pay close attention to the faces of the girls, especially their noses. These features suggest subtle inconsistencies, indicating that the noses may not be real and were likely crafted by artificial intelligence.

scroll for more

Fake

This is a synthetic image produced by AI. Take a look at the girls' faces, especially their noses. They look a bit off, which might be a clue that the noses aren’t real but were created by AI.

scroll for more

Real

This is a real image.

Real

This is a real image.