X tightens AI safeguards, restricts Grok from generating objectionable images of real people
PTC Web Desk: Microblogging platform X (formerly Twitter) has introduced new technical safeguards to prevent its artificial intelligence chatbot Grok from generating or editing objectionable images of real individuals, following widespread backlash over the misuse of AI-generated deepfake content.
The company announced that the restrictions will apply to all users, and access to Grok’s image-generation and editing features will now be limited only to paid subscribers, a move aimed at ensuring greater accountability and curbing misuse. X has stated that this step is intended to discourage the creation and circulation of non-consensual and sexually explicit AI-generated content.
In addition, X has implemented geo-blocking measures in regions where such content is illegal. The decision comes after India’s Ministry of Electronics and Information Technology (MeitY) raised strong objections to the circulation of obscene content generated through Grok and warned the platform of strict legal action.
Following the government’s intervention, X acknowledged lapses in enforcement and confirmed that it has removed nearly 3,500 pieces of objectionable content and deleted more than 600 accounts involved in violating platform and legal norms. The Indian government had earlier cautioned that failure to comply with provisions of the Information Technology Act could result in the platform losing its legal safe-harbour protections.
In response, X has assured full compliance with Indian laws and reiterated its zero-tolerance policy towards sexual exploitation, deepfake abuse and the sharing of obscene content without consent.
- With inputs from agencies
