The World of AI Deepfakes & AI Misinformation

500K+ deepfakes online
(2023 estimate)
96% target women for
non-consensual content
$1B+ annual harm cost
estimated by researchers

What is a DeepFake?

DeepFakes are essentially AI generated images made to look like real people. This can be used to impersonate
someone or depict them doing things theey wouldn't normally do. DeepFakes also aren't limited to just images. You
can make videos and voice deepfakes as well which can allow you to for example, make the president run away from
the police, or put someone in any clothing of your choice. Unfortunately that second example is more real than you would think.

In the News

The effects of deepfakes are wide-ranging and can cause major harm if left unchecked.

Non-consesual Images

One of the major examples how DeepFakes harm people is through creating images of people doing things they would never actually do. For example, there have being news reports of students using deepfakes to make staff members say very inapropriate things. When this happened to a principal, it took weeks to clear his name. Politians aren't immune to this either. DeepFakes can be used to make politicians say anything, leading to the potential for massive misinformation campaigns and the spread of confusion. Deepfakes are even used to create explicit non-consensual sexual images of women. This is a form of sexual harassment that has become incresingly common and increasingly hard to stop as these images are spread across the internet leading to major harm done to the women who are faked. Unfortunately, even young girls get caught up in these terrible actions leading to major damage and harm being done to thier mental health and self confidence.

The Scam Revolutions

text 2

Significance

significance

Legislation

legislation

Platform Moderation

moderation

Model Guardrails

guardrails

Take the quiz!

Can you tell the difference between real and AI-generated faces?

interactive game goes here

Sources

  1. Deeptrace Labs — State of Deepfakes Report
  2. WIRED — The Deepfake Problem is Getting Worse
  3. Congress.gov — DEFIANCE Act 2024
  4. MIT Technology Review — AI Detection Tools
  5. thispersondoesnotexist.com (StyleGAN2)
  6. deepware.ai