India orders Musk’s X to fix Grok over “obscene” AI content

India has ordered Elon Musk’s X to marque contiguous method and procedural changes to its AI chatbot Grok aft users and lawmakers flagged the procreation of “obscene” content, including AI-altered images of women created utilizing the tool.

On Friday, India’s IT ministry issued the bid directing Musk’s X to instrumentality corrective enactment connected Grok, including restricting the procreation of contented involving “nudity, sexualization, sexually explicit, oregon different unlawful” material. The ministry besides gave the societal media level 72 hours to taxable an action-taken study detailing the steps it has taken to forestall the hosting oregon dissemination of contented deemed “obscene, pornographic, vulgar, indecent, sexually explicit, pedophilic, oregon different prohibited nether law.”

The order, reviewed by TechCrunch, warned that nonaccomplishment to comply could jeopardize X’s “safe harbor” protections — ineligible immunity from liability for user-generated contented nether Indian law.

India’s determination follows concerns raised by users who shared examples of Grok being prompted to change images of individuals — chiefly women — to marque them look to beryllium wearing bikinis, prompting a formal complaint from Indian parliamentarian Priyanka Chaturvedi. Separately, caller reports flagged instances successful which the AI chatbot generated sexualized images involving minors, an contented X acknowledged earlier connected Friday was caused by lapses successful safeguards. Those images were aboriginal taken down.

However, images generated utilizing Grok that made women look to beryllium wearing bikinis done AI alteration remained accessible connected X astatine the clip of publication, TechCrunch found.

The latest bid comes days aft the Indian IT ministry issued a broader advisory connected Monday, which was besides reviewed by TechCrunch, to societal media platforms, reminding them that compliance with section laws governing obscene and sexually explicit contented is simply a prerequisite for retaining ineligible immunity from liability for user-generated material. The advisory urged companies to fortify interior safeguards and warned that nonaccomplishment to bash truthful could invitation ineligible enactment nether India’s IT and transgression laws.

“It is reiterated that non-compliance with the supra requirements shall beryllium viewed earnestly and whitethorn effect successful strict ineligible consequences against your platform, its liable officers and the users connected the level who interruption the law, without immoderate further notice,” the bid warned.

The Indian authorities said non-compliance could pb to enactment against X nether India’s IT instrumentality and transgression statutes.

India, 1 of the world’s biggest integer markets, has emerged arsenic a captious trial lawsuit for however acold governments are consenting to spell successful holding platforms liable for AI-generated content. Any tightening of enforcement successful the state could person ripple effects for planetary exertion companies operating crossed aggregate jurisdictions.

The bid comes arsenic Musk’s X continues to situation aspects of India’s content regularisation rules successful court, arguing that national authorities takedown powers hazard overreach, adjacent arsenic the level has complied with a bulk of blocking directives. At the aforesaid time, Grok has been progressively utilized by X users for real-time fact-checking and commentary connected quality events, making its outputs much disposable — and much politically delicate — than those of standalone AI tools.

X and xAI did not instantly respond to requests for remark connected the Indian government’s order.