The proliferation of nonconsensual, sexualized deepfakes has grown into a systemic problem across the technology industry, extending far beyond any single platform. U.S. lawmakers are now pressing major social media and technology companies to demonstrate how they are addressing the spread of AI-generated sexual imagery that depicts real people without consent.
In a letter sent to executives at X, Meta, Alphabet, Snap, Reddit, and TikTok, a group of U.S. senators requested detailed evidence that these companies have “robust protections and policies” in place to combat sexualized deepfakes. The lawmakers also demanded that the companies preserve all documents and internal records related to the creation, detection, moderation, and monetization of sexualized AI-generated images, as well as any policies governing such content.
The letter was sent just hours after X announced updates to its Grok chatbot, including restrictions that prohibit editing images of real people into revealing clothing and limiting image generation and editing features to paying subscribers. X and xAI operate under the same corporate umbrella, and Grok has recently come under intense scrutiny following media reports documenting how easily it produced sexualized and nude images of women and minors.
Citing those reports, the senators argued that existing platform safeguards designed to prevent nonconsensual sexual imagery are proving inadequate. While many companies publicly claim to block explicit content and prohibit nonconsensual intimate imagery, lawmakers said real-world examples show that these protections are frequently bypassed or fail outright.
“We recognize that many companies maintain policies against non-consensual intimate imagery and sexual exploitation, and that many AI systems claim to block explicit pornography,” the letter states. “In practice, however, users are finding ways around these guardrails. Or these guardrails are failing.”
Although Grok and X have drawn particular criticism, the senators emphasized that the issue spans the broader digital ecosystem. Sexualized deepfakes first gained widespread attention in 2018, when a subreddit hosting synthetic pornographic videos of celebrities went viral before Reddit removed it. Since then, such content has multiplied across platforms including TikTok and YouTube, often circulating widely even if it originates elsewhere.
Meta has faced repeated scrutiny over explicit AI-generated images of female public figures, including cases flagged by the company’s own Oversight Board. The company has also hosted advertisements for so-called “nudify” apps on its platforms, though it later pursued legal action against one such service, CrushAI. Snapchat has been linked to multiple incidents involving minors sharing deepfakes of classmates, while Telegram—though not named in the senators’ letter—has become notorious for bots that digitally undress images of women.
In response to the lawmakers’ inquiry, X referred to its recent changes to Grok. Reddit issued a statement emphasizing that it prohibits nonconsensual intimate media in all forms, including AI-generated or manipulated imagery, and said it does not provide tools capable of creating such content. The company added that it bans the solicitation of nonconsensual imagery, links to nudify apps, and instructions for producing such material. Alphabet, Meta, Snap, and TikTok did not immediately respond to requests for comment.
The senators’ letter outlines an extensive list of disclosures the companies are expected to provide, including definitions of deepfake content and nonconsensual intimate imagery; explanations of how policies apply to altered or partially clothed images and so-called “virtual undressing”; details on moderation guidance and enforcement practices; descriptions of safeguards built into AI tools; mechanisms for detecting deepfakes and preventing re-uploads; policies preventing users or platforms from monetizing such content; enforcement powers under terms of service; and procedures for notifying victims.
The letter is signed by Senators Lisa Blunt Rochester, Tammy Baldwin, Richard Blumenthal, Kirsten Gillibrand, Mark Kelly, Ben Ray Luján, Brian Schatz, and Adam Schiff.
The congressional action follows comments made earlier this week by xAI owner Elon Musk, who said he was “not aware of any naked underage images generated by Grok.” Shortly afterward, California’s attorney general opened an investigation into xAI’s chatbot amid growing international pressure over the lack of effective safeguards that allowed such content to be generated. While xAI has stated that it removes illegal material from X, including child sexual abuse material and nonconsensual nudity, neither the company nor Musk has addressed why Grok was capable of generating the content in the first place.
The problem extends beyond sexualized imagery alone. Even AI tools that do not explicitly enable “undressing” features often allow the creation of realistic deepfakes. Reports have indicated that OpenAI’s Sora 2 permitted the generation of explicit videos involving children, that Google’s Nano Banana model produced violent political imagery, and that racist AI-generated videos using Google’s video tools have accumulated millions of views on social platforms.
The situation becomes more complex with the involvement of Chinese image and video generation tools, many of which offer advanced face, voice, and video manipulation features. Content produced by these systems has increasingly circulated on Western platforms. While China enforces stricter labeling requirements for synthetic media, the United States lacks comparable federal standards, leaving enforcement largely in the hands of individual platforms with inconsistent results.
Although federal lawmakers have passed legislation aimed at curbing deepfake pornography, its effectiveness has been limited. The Take It Down Act, enacted in May, criminalizes the creation and distribution of nonconsensual sexual imagery but places much of the enforcement burden on individual users rather than on platforms or AI tool providers. In response, several states are pursuing their own measures. This week, New York Governor Kathy Hochul proposed legislation requiring AI-generated content to be clearly labeled and banning nonconsensual deepfakes during defined periods leading up to elections.
As generative AI tools become more powerful and accessible, lawmakers argue that voluntary platform policies are no longer sufficient, and that stronger accountability mechanisms may be necessary to prevent widespread harm.