The proliferation of advanced AI tools has industrialized the abuse of women online, with millions of people around the world creating and sharing deepfake nudes on the secure messaging app Telegram, a Guardian analysis has found.
The Guardian has identified at least 150 Telegram channels – large encrypted group chats popular for secure communication – that appear to have users in many countries, from the UK to Brazil, China to Nigeria, Russia to India. Some offer “nude” photos and videos for a fee, allowing users to upload a photo of any woman and the AI will generate a video of her performing a sex act. Additionally, many sites offer feeds of images of celebrities, social media influencers, and everyday women rendered nude or forced into sexual acts by AI. Followers also use this channel to share tips about the deepfake tools available.
Telegram channels dedicated to distributing non-consensual nude images of women have long existed, but with the proliferation of AI tools, anyone can become the subject of graphic sexual content that can be viewed by millions in an instant.
On a Russian-language Telegram channel promoting deepfake “blogger leaks” and “celebrity leaks,” a post about an AI-denuding Telegram bot promised “a neural network that doesn’t know the word ‘no.'”
“Choose position, shape, location. Do with her everything you can’t do in real life,” it read.
On the Chinese-language Telegram channel, which has about 25,000 subscribers, men shared videos of their “first love” or “girlfriend’s best friend” being forced to strip using AI.
A web of Telegram channels targeting users in Nigeria is spreading deepfakes along with hundreds of stolen nudes and intimate images.
Telegram is a secure messaging app that allows users to create groups and channels and broadcast content to an unlimited number of contacts. The platform’s terms of service prohibit users from posting “illegal pornographic content” on “publicly viewable” channels or bots or “participating in activities that are considered illegal in the majority of countries.”
A review of data from Telemetr.io, an independent analytics and database service that includes an index of such channels, shows that Telegram has shut down a large number of exposed channels.
Telegram told the Guardian that deepfake porn and the tools to create it are expressly prohibited in its terms of service, adding: “Such content is routinely removed as it is discovered. Moderators, empowered with custom AI tools, actively monitor public parts of the platform and accept reports to remove content that violates our terms of service, including encouraging the creation of deepfake porn.”
Telegram said in a statement that it removed more than 952,000 pieces of harmful content in 2025.
The use of AI tools to create sexual deepfakes and shame women has exploded into public debate in recent weeks after Elon Musk’s social media platform
In response to the resulting outrage, Musk’s artificial intelligence company xAI announced it would no longer allow Grok to edit photos of real people into bikinis. British media regulator Ofcom also announced an investigation into X.
But there is a treasure trove of forums, websites and apps, including Telegram, that give millions of people easy access to graphic, non-consensual content, and allow them to generate and share this content on demand, without the knowledge of the women being violated. Dozens of nudity apps are available on the Google Play Store and Apple App Store, with a combined total of 705 million downloads, according to a report released Tuesday by the Tech Transparency Project.
An Apple spokesperson said the company has removed 28 of the 47 nude apps identified by the Tech Transparency Project in its investigation, while a Google spokesperson said “most apps” on its services have been suspended and an investigation is ongoing.
Anne Cranen, an expert on gender-based violence at the London-based Institute for Strategic Dialogue, said the Telegram channel was the mainstay of a broader internet ecosystem dedicated to the creation and distribution of non-consensual intimate images.
This allows users to circumvent the control of big platforms like Google and share tips on how to circumvent safeguards that prevent AI models from producing this content. But “the dissemination and celebration of this material is another part,” she said. “The spreading of it to other men, the bragging about it, and the celebratory aspect is also very important. This really shows the root of that misogyny. They’re trying to punish women or silence women.”
Last year, Meta shut down an Italian Facebook group where men shared intimate images of their partners and unsuspecting women. before the group is deleted. mia mogli (meaning “My Wife”) had approximately 32,000 members.
However, investigative newsletter Indicator revealed that Meta has been unable to stop the flow of ads for AI disabling tools on its platform, identifying at least 4,431 disabling ads across its platforms since December 4 last year, some of which appear to be fraudulent. A Meta spokesperson said the company removes ads that violate its policies.
AI tools have intensified online violence against women globally, allowing almost anyone to create and share abusive images. In many jurisdictions, including large parts of the Global South, there are few legal avenues to hold perpetrators to account. According to 2024 World Bank data, less than 40% of countries have laws protecting women and girls from cyberharassment and cyberstalking. The United Nations estimates that 1.8 billion women and girls still lack legal protection from online harassment and other forms of abuse facilitated by technology.
Lack of regulation is just one reason women and girls in low-income countries are particularly vulnerable, campaigners say. Issues such as low digital literacy and poverty can increase risk. Ugochi Ihe, an associate at TechHer, a Nigeria-based organization that encourages women and girls to learn and use technology, said she has come across cases where women borrowing money from loan apps are “falling victim to extortion by unscrupulous men using AI. The abuses are getting more creative by the day.”
The real-life effects of digital abuse are devastating, including mental health difficulties, isolation, and job loss.
“These events always end up destroying a girl’s life,” said Mercy Mutemi, a Kenya-based lawyer who represents four victims of deepfake abuse. She said some of her clients have been refused jobs or had disciplinary hearings at school because of deepfake images that were distributed without their consent.
Ihe said her organization fielded complaints from women who were blackmailed and ostracized by their families with nude and intimate images obtained from Telegram channels.
“Once it’s leaked, you can’t regain your dignity or identity. Even if the perpetrator comes and says, ‘Oh, that was a deepfake,’ you don’t know how many people saw it. The reputational damage is irreversible.”
