Editorial illustration for Grok still hosts sexualized deepfakes of famous women; Musk added undress button
Grok still hosts sexualized deepfakes of famous women;...
Months after promising to crack down, Elon Musk's Grok chatbot is still generating sexualized deepfakes of women without their consent. That’s according to new reporting from WIRED. The explicit images, which include fabricated depictions of celebrities like Taylor Swift, continue to be hosted on the platform.
Then, last week, an “undress” button briefly appeared. It was a function that explicitly invited users to strip clothing from images of real people.
Elon Musk’s Grok chatbot is apparently still being used to produce and host nonconsensual explicit images and videos of women, months after Musk’s artificial intelligence firm xAI said it would introduce restrictions to stop the creation of potentially harmful sexualized deepfakes.
That "undress" feature vanished within a day. Its fleeting release, however, followed a pattern critics find alarming. AI ethics researchers argue the combination—a persistent deepfake problem plus a tool for generating more—reveals a product that enables harassment.
xAI has offered no details on the technical guardrails it pledged back in March. The result is stark: Grok still produces the very nonconsensual deepfakes it claimed it would restrict.
Related Reading
Business & Startups Nordic pilot adds Gemini for Education, NotebookLM to boost AI literacy
Business & Startups Kling launches Video O1, all-in-one model with MVL bridge using transformer
Business & Startups DeepSeek Seeks More Capital Weeks After USD 7B Funding Round
Related Topic Anthropic sues Trump as xAI restarts AI coder push, Iran releases AI fakes
Open Source
Related Topic Jeffrey Epstein Consulted Elon Musk Ally on Potential Tesla Privatization
Open Source
From Archives xAI sues after firing who warned of Grok safety; he led Scale AI safety work
From Archives Anthropic launches Fable 5, blocks cybersecurity, biology, chemistry queries
From Archives xAI used Anthropic’s Claude via personal accounts after access revoked for months
LLMs & Generative AI
From Archives xAI launches Grok Imagine 1.5, adding 720p text‑prompted image‑to‑video
Open Source
Common Questions Answered
What nonconsensual deepfakes has Grok been generating despite Musk's promises to crack down?
According to WIRED reporting, Grok has been generating sexualized deepfakes of famous women without their consent, including fabricated explicit images of celebrities like Taylor Swift. These nonconsensual deepfakes continue to be hosted on the platform months after Musk promised to address the issue.
What was the 'undress' button that briefly appeared on Grok and why did it raise concerns?
The 'undress' button was a function that explicitly invited users to remove clothing from images of real people, effectively enabling the creation of nonconsensual sexualized content. The feature vanished within a day but demonstrated a pattern that AI ethics researchers argue reveals a product designed to facilitate harassment.
What technical guardrails did xAI pledge to implement for Grok and have they been disclosed?
xAI pledged back in March to implement technical guardrails to restrict nonconsensual deepfakes on Grok. However, the company has offered no details on what these safeguards actually are or how they function, leaving critics skeptical about their effectiveness.
How do AI ethics researchers characterize the combination of Grok's deepfake problem and the undress tool?
AI ethics researchers argue that the combination of persistent deepfake generation capabilities plus an explicit tool for stripping clothing from images reveals a product that enables harassment of real people. This pattern demonstrates a failure to implement promised protections against nonconsensual sexual content.
Further Reading
- A New Form of Gendered Violence: Elon Musk's Grok — Human Rights Research
- The Grok Scandal Forces Reckoning on AI, Consent and Gendered Harms — CIGI Online
- Grok AI Deepfake — Wallace Miller
- Grok sexual deepfake scandal — Wikipedia
- Non-consensual Grok deepfakes endanger women — Hacker News
