US enforces law to crack down on sexual deepfakes
Sign up now: Get ST's newsletters delivered to your inbox
A wave of AI porn scandals has been reported at schools across the US, with hundreds of teenagers targeted by their own classmates.
PHOTO: PEXELS
WASHINGTON – The US on May 19 began enforcing a law requiring tech platforms to remove sexual deepfakes and other non-consensual intimate imagery, but experts warned of shortcomings and raised concerns about online censorship.
US President Donald Trump signed the Take It Down Act in 2025, criminalising the online distribution of non-consensual sexual imagery, which is often created using cheap and widely available artificial intelligence tools.
The Federal Trade Commission (FTC) said that starting on May 19, tech platforms were required to establish a process allowing victims to request the removal of such content and must take it down within 48 hours of receiving a valid request, or face penalties.
“We stand ready to monitor compliance, investigate violations and enforce the Take It Down Act,” FTC chairman Andrew Ferguson said after sending letters to more than a dozen tech firms – including Meta, TikTok, X and Snapchat – ahead of the enforcement.
“Protecting the vulnerable – especially children – from this harmful abuse is a top priority for this agency and this administration.”
In a post on May 18, X’s Safety account said that there is “no place in our society for predators to share intimate photos and videos of others without their consent”.
“X has zero tolerance for non-consensual intimate images, unwanted sexual content or any kind of exploitative behaviour.”
‘Shoot first, ask questions never’
The assurance comes after Mr Elon Musk’s AI tool Grok faced international backlash earlier in 2026 for virtually undressing women and minors without their consent. Researchers said Grok generated an estimated three million sexualised images in a matter of days.
Still, some experts warn that the Take It Down Act is far from a perfect solution.
Ms Riana Pfefferkorn, a policy fellow at the Stanford Institute for Human-Centered Artificial Intelligence, said the law creates a “shoot first, ask questions never” dynamic.
The incentive structure is in the “direction of just remove it, remove it, remove it”, she was quoted as saying in Indicator, a newsletter focused on investigating digital deception.
Ms Pfefferkorn also expressed concern that the law could be weaponised against trans people, sex workers whose content is consensual, and political speech that the administration disagrees with.
The law’s takedown provision has also alarmed other free speech advocates, who warn it could encourage tech firms to zealously moderate non-offending content to reduce risk.
An online boom in non-consensual deepfakes is outpacing efforts to regulate the technology worldwide because of the proliferation of AI tools, including “nudification” apps, experts say.
While high-profile politicians and celebrities, including singer Taylor Swift, have been victims of deepfake porn, researchers say women not in the public eye are equally vulnerable.
A wave of AI porn scandals has been reported at schools across US states – from California to New Jersey – with hundreds of teenagers targeted by their own classmates.
Such non-consensual imagery can lead to harassment, bullying or blackmail, sometimes causing devastating mental health consequences, experts warn. AFP
