Researchers find a Telegram bot that takes a person's photo and turns it into a fake nude; by July, 100K+ images of women had been transformed, shared publicly
Bots have been used to create more than 100,000 images — Researchers have discovered a “deepfake ecosystem” …
Context & Ripple Effects
When researchers surfaced this bot, they exposed more than a single tool: a service that kept operating a month later despite public disclosure, running inside Telegram's public channel structure where transformed images of women were openly shared. Within days, Italy opened an investigation into the fake-nude bots, making this one of the first state responses to consumer-grade nonconsensual imagery tools.
The discovery has since proven to be an early snapshot of an industry rather than an isolated abuse case: coverage traced growth from this one bot to 50+ bots claiming 4M+ combined monthly users by 2024, and by 2026 the ecosystem had moved into high-quality video, drawing 1.4M+ accounts across 39 creation bots. The through-line is that disclosure and one national probe did not constrain distribution.
First-order effects
- Women whose photos were processed — 100K+ images by July alone — became victims of publicly shared nonconsensual sexualized imagery with no removal mechanism named in the reporting.
- Telegram faced immediate pressure to police its public channels, yet researchers found the service continued operating largely unchanged after disclosure.
Second-order effects
- National regulators, starting with Italy's investigation, were forced to treat bot-based image abuse as an enforcement question for the hosting platform rather than for individual bad actors.
- Each documented failure to act raised the cost of the next discovery: follow-up research escalated from cataloging one bot to auditing the entire Telegram deepfake economy, keeping sustained pressure on the platform.
Third-order effects
- If the trajectory from one bot to dozens holds, nonconsensual synthetic imagery becomes an industrialized category on messaging platforms, shifting accountability debates toward platform-level obligations and regulation of hosts like Telegram.
- The move from images to video signals that detection and provenance tooling built for stills will lag the formats actually circulating, pushing the burden onto trust infrastructure for synthetic media generally.
The trend: Consumer-grade deepfake tools are scaling from single novelty bots into a persistent commercial ecosystem on messaging platforms, outpacing both moderation and early regulatory probes.