AI fueles a surge in online abuse against women in India: ‘It felt like someone had taken control of how I exist visually’

It was just the typical New Year’s Eve selfie shared on social media, but the image soon took a dark turn, and its creator, Ruchi Kokcha, watched as AI-generated versions of the photograph depicting her naked spread like wildfire across the internet in her country, India, without her permission and with nothing she could do to stop it.

“It was digital sexual abuse,” Kokcha tells EL PAÍS in New Delhi, explaining that anonymous users had run the image through Grok, a generative AI chatbot designed by Elon Musk’s xAI. “When someone uses AI to create fake sexual images of a person without their consent, it is a violation of basic human rights. Just because it is online doesn’t make it any less of a crime.”

Kokcha’s experience is part of a rapidly expanding problem across India’s digital landscape, where generative artificial intelligence tools are being used to produce convincing deepfakes, sexualized images, and voice clones in minutes, abuses that disproportionately target women.

“The problem reflects deeper issues within society. Women are disproportionately targeted, not only in India, but across the world,” says Nikhil Pahwa, founder and editor of MediaNama, a publication focused on digital policy in India.

Under India’s Information Technology Rules and the proposed Digital India Act, authorities could impose fines of up to 500 million rupees (about $5.4 million) on platforms that fail to prevent the spread of harmful content. Pahwa says laws are only as effective as their implementation.

In India, the controversy reached a fever pitch on Jan. 2, when India’s Ministry of Electronics and Information Technology issued a notice to X, condemning it for allowing Grok to “generate and circulate obscene, sexually explicit and derogatory content, particularly targeting women and children,” and demanding that it remove all unlawful content.

After India’s complaint, X said it had blocked about 3,500 pieces of content and removed more than 600 accounts tied to sexualized images created with the tool. But the damage was already done. In the lightning-quick era of social media, where explicit content can spread in seconds, experts say that platform moderators can barely keep up, and the glacial pace of the courts offers little legal remedy when reputations and privacy are being destroyed in real time.

Online auction of women

In India, developments in generative AI have fuelled increasingly toxic forms of harassment. In 2021, an app called Sulli Deals appeared on GitHub, mock-auctioning photos of hundreds of women, including journalists, students, and activists, without consent. The goal was to humiliate them. When a near-identical platform, Bulli Bai, resurfaced the harassment in January 2022, arrests came only after months of public pressure and media scrutiny.

More recent cases include a sexually explicit 19-minute AI-generated deepfake of YouTuber Payal Gaming that went viral across Telegram and Facebook, and morphed images of actor Girija Oak, who said she’s horrified by the idea that her 12-year-old son may find them.

“Women have to constantly calculate risk before simply existing online,” says Kokcha. “The internet was not supposed to be this. At the very least, we need quicker action from platforms, clearer laws, and a shift in how seriously this is taken.”

Kokcha’s case coincided with a new image-generation and editing feature added to Grok in December 2025 that allowed users to upload photos and alter them using simple text prompts. Within days, the feature was being used to generate an avalanche of sexualized images of women and minors without their consent. The Center for Countering Digital Hate found that Grok generated approximately three million sexualized images, including 23,000 that appeared to depict children, in just 11 days, between December 29, 2025, and January 8, 2026.

The xAI tool has sparked legal action in the United States, the Netherlands and the United Kingdom and led EU countries in March to agree to call for a ban on AI systems that allow the creation of sexualized images without consent and the generation of child sexual abuse material.

For Kokcha, AI-generated harassment is even more sinister than the online abuse women have long grown accustomed to. “Trolling is horrible, but you can still somewhat separate yourself from it, as it’s just words, opinions, or abuse,” she says. “This felt like someone had taken control of how I exist visually. That line between ‘this is me’ and ‘this isn’t me’ becomes very blurry, and that’s deeply unsettling.”

Self-censorship

Amid the criticism, xAI restricted image generation to paying subscribers and geographically blocked the ability to generate nude images of people in jurisdictions where such content is illegal. However, experts say that explicit images are still being produced and that the problem is far from resolved. Other platforms, such as Meta, have introduced measures to reduce obscene content. But X’s Grok lacks comparable safeguards.

Faced with this situation, self-censorship is already becoming evident. Women are speaking and sharing less because they perceive the online environment as unsafe. Some have stopped posting photos of themselves, while others avoid using profile pictures altogether, according to Pahwa.

“A single image can be turned into a deepfake video or sexualised content and then replicated at scale,” says Apar Gupta, founding director of the Internet Freedom Foundation. “For women, this shifts harassment from text-based insults to ‘evidence-like’ media that can trigger stigma, job harm, and coercion.”

Priyanka Chaturvedi, a member of India’s parliament, was among the first political leaders to formally raise concerns about the misuse of AI tools on social media. She wrote to the Union Minister for Electronics and Information Technology, describing the proliferation of deepfakes of women — with clearly sexual content — circulating on social media.

“More than trolling, it is the misrepresentation of women that is deeply concerning,” Chaturvedi tells El País. “I have seen several AI-generated videos of myself that are highly objectionable.”

The lawmaker acknowledged steps taken by the government, including takedown mechanisms and regulatory safeguards, but warned that the pace of AI development is outrunning enforcement. Chaturvedi said that deepfakes can appear so convincing that it is difficult for any government to impose effective safeguards. That’s why she called for greater international cooperation, stronger accountability, and faster mechanisms to remove harmful content.