"You are not powerless": Melissa Hutchins on guarding your face from deepfakes
At TEDxUofW, the cyberstalking survivor sets out a three-layer shield, from privacy settings to a family safe word to new federal law.
Why this story? The video opened at a pace of 55,401 views a day, 105 times a normal first day on this channel (529).
How to protect your identity from AI deepfakes | Melissa Hutchins | TEDxUofW
Three to five photos. That, Melissa Hutchins tells her audience at TEDxUofW, is all someone now needs: "Today, three to five ordinary photos taken from someone's social media are all it takes."
She speaks from experience. In her early twenties, Hutchins endured a severe cyberstalking campaign that lasted five years and ended in a federal trial involving the US Secret Service. She recalls a federal judge telling her, "I can't give you the level of protection for the length of time that you deserve." Near the end of the talk, she describes how the trial concluded in 2024 with a nine-year federal prison sentence.
The middle of the talk is about the wider problem. Hutchins explains non-consensual intimate imagery and how easy AI tools have made it to produce. She shows the audience four portraits and asks which faces are real, then tells them that people pick out the AI-generated ones barely half the time. "96% of deepfake videos online are sexually explicit, non-consensually generated, and 99% are targeting women and girls," she says. She urges parents to think before posting their children's photos publicly, offering a question to ask first: "Does this need to be public, or can it be shared with only the people that I trust most?"
Then she turns to what people can do. Her "three-layer shield" starts with locking down privacy settings. Next comes a family safe word, agreed offline, as a check against scams that imitate a familiar voice. The third layer is knowing the law, including the Take It Down Act. A final section asks for a change in culture: "If somebody tells you this happened to them, believe them." Hutchins has since founded Certifi AI, a company working against malicious deepfakes and digital impersonation.
The talk passed 36,000 views in its first 11 hours. A typical new talk on the channel draws a few hundred views in a full day, and talks posted the same day, including two on medicine from TEDxBoston Salon, stayed in that range. Under her video, viewers ask how to keep their identity safe and where to turn for help. One adds a practical refinement to her safe word: never write it down in texts or group chats.
Published on Trari
Where next?
- Channel
- TEDx Talks
- Subscribers
- 44,700,000
- Channel views, last 7 days
- 11,883,571
- Channel views, pace per day
- 601,636 (on a normal day: 1,949,429)
- Video published
- October 12, 2026
- Video views so far
- 36,471 (less than a day online)
- Video views, pace per day
- 55,401
The numbers
One talk out of a crowded upload day pulled far ahead of how new talks usually start on this channel.



