X to Stop Grok AI from Undressing Images of Real People After Backlash
Elon Musk’s AI model, Grok, will no longer edit photos of real individuals to depict them in revealing clothing, following widespread concern over sexualized AI deepfakes.
New Measures Implemented for Grok AI
– Preventive Action: We have implemented technological measures to prevent the Grok account from allowing the editing of images of real people in revealing clothing, stated a recent announcement on X.
– Government Response: The UK government expressed vindication after Prime Minister Sir Keir Starmer urged X to manage its AI tool more effectively. This announcement came just hours after California’s top prosecutor revealed that the state was investigating the proliferation of sexualized AI deepfakes, including those involving children.
– Geoblocking Measures: We now geoblock the ability of all users to generate images of real people in bikinis, underwear, and similar attire via the Grok account and in Grok in X in those jurisdictions where it’s illegal, X confirmed in a statement on Wednesday.
– Paid User Limitation: The company emphasized that only paid users would retain the ability to edit images using Grok, adding that this would ensure accountability for those who might misuse the AI tool in violation of laws or X’s policies.
Addressing Concerns Over Grok AI Deepfakes
– Content Restrictions: With NSFW (not safe for work) settings enabled, Grok will now permit upper body nudity of imaginary adult humans (not real ones), according to Musk. He noted that this aligns with standards seen in R-rated films, which may differ according to local laws.
– Global Criticism: Leaders worldwide have criticized Grok’s image-editing functionality. Recently, Malaysia and Indonesia became the first nations to prohibit the Grok AI tool after users reported that photos were altered to create explicit images without consent.
– Regulatory Investigation: Britain’s media regulator, Ofcom, stated Monday that it would investigate whether X had violated UK law regarding sexual images. If found guilty, X could face fines up to 10% of its worldwide revenue or £18 million, whichever is greater.
Prime Minister’s Stance and Future Actions
– Potential Loss of Self-Regulation: Sir Keir warned that X could lose the right to self-regulate amid rising backlash over the AI-generated images. However, he later expressed approval of X’s actions to address the issue.
– Legislative Measures: The Prime Minister indicated he would take the necessary measures to strengthen legislation if X fails to act adequately. He previously condemned the images produced by Grok as disgusting and shameful, labeling the decision to monetize this service as horrific.
Ongoing Challenges and Concerns
– Grok Safeguards Delayed: California Attorney General Rob Bonta highlighted that material depicting women and children in compromising situations has been used to harass individuals online.
– Need for Clarity: Riana Pfefferkorn, a policy researcher, expressed surprise that X took so long to implement safeguards for Grok, arguing that the AI’s editing features should have been removed swiftly. She raised questions regarding how X will enforce its new policies—particularly how the AI will identify real individuals and respond to violations.
– Musk’s Role Questioned: Pfefferkorn noted that Musk’s public antics, such as sharing an AI-generated image of Sir Keir Starmer in a bikini, undermine the seriousness of the situation.
As the fallout continues, it remains crucial for X and Musk to establish effective guidelines and enforcement mechanisms to ensure the responsible use of AI technologies like Grok.