One month after its discovery, researchers say a Telegram bot that makes fake nudes of women continues to operate and the service has done little to stop it
A bot automated deepfake abuse online. Researchers say their warnings have been ignored and Telegram has failed to remove it
Context & Ripple Effects
When researchers first exposed the bot in October 2020, they found more than 100K images of women had already been transformed and shared across public Telegram channels (the initial findings) — and within days, Italy opened an investigation into these fake-nude bots. This follow-up is the accountability check: a month on, the bot still runs and Telegram has done little despite direct warnings.
What makes the inaction consequential is what came after. Later reviews of Telegram communities found 50+ bots claiming to create explicit nonconsensual content with 4M+ claimed monthly users, and the ecosystem has since expanded into high-quality video generation with 1.4M+ accounts signed up to creation bots — suggesting the moderation gap documented here never closed.
First-order effects
- Women targeted by the bot remain exposed: any photo of them can still be turned into a fake nude, and the images already circulating in Telegram's public channels stay up alongside the tool that produced them.
- Telegram's own reporting channels are shown to be ineffective for this category — researchers who escalated directly were ignored, meaning victims have no working removal path on the platform.
Second-order effects
- Italy's investigation becomes the template for other regulators: with voluntary takedowns failing, national authorities are pushed toward formal probes and enforcement against Telegram itself rather than individual bot operators.
- The unremoved bot signals to copycat developers that Telegram carries minimal enforcement risk, helping explain why reviews later found dozens of competing abuse bots on the same platform.
Third-order effects
- If the pattern holds, explicit deepfake abuse consolidates around messaging platforms as distribution-and-creation infrastructure, shifting the burden onto regulators and victim-side remedies because platforms won't act voluntarily.
- Sustained platform inaction strengthens the case for legal mandates — like the investigative path Italy chose — that hold host platforms liable for synthetic nonconsensual content rather than treating each bot as a discrete case.
The trend: Nonconsensual deepfake abuse is scaling from single viral bots into a durable creator ecosystem on messaging platforms, with platform moderation lagging far behind and regulators stepping into the vacuum.