How Telegram Became a Hub for Re-Uploaded Creator Content

Telegram's combination of large-membership public channels, searchable bots, and light content moderation has made it a persistent destination for re-uploaded creator content, from pirated video to non-consensually shared intimate images. The numbers below, drawn from academic research, investigative journalism, and child-safety organizations, sketch out how large that ecosystem actually is.

1. The scale of Telegram's piracy ecosystem

Researchers have only recently begun measuring Telegram's piracy footprint at scale, and the first large academic study of it found an ecosystem far bigger than anecdotal reporting suggested.

A first-of-its-kind academic study analyzing 1,057 Telegram channels and roughly 209,000 posts between December 2023 and January 2026 identified 19,033 unique pirated movie and TV titles from 3,941 production companies.^[1] The same study estimated a lower-bound financial loss to rightsholders of $17.49 billion, tied to 4.85 billion accumulated views, and mapped an internal redirection network of 2,312 nodes — 1,875 channels and 437 bots — linked by 3,843 directed edges that quietly funnel viewers between channels.^[1]

Separate detection work backs up how much churn exists inside that same network. An AI piracy-detection tool called Anti-RIP scanned 249,133 new Telegram channels over roughly two months and flagged 802 piracy channels, 299 connected channels, and 108 bots.^[2] Reporting those flagged channels to Telegram and to rightsholders had a measurable real-world effect: 524 of 1,101 reported channels went inaccessible within two weeks, and over a 61-day window 524 channels and 71 bots were ultimately removed.^[2]

2. Nudify and deepfake bots operating openly

Alongside pirated video, Telegram has become a home for bots that generate non-consensual intimate imagery, often built directly on top of the platform's bot API.

A WIRED investigation identified at least 50 AI "nudify" bots operating on Telegram with a combined 4 million-plus monthly users, backed by at least 25 channels with more than 3 million combined members.^[3] Beyond the two largest bots, 14 additional bots each individually surpassed 100,000 monthly users.^[4] After WIRED contacted Telegram about the investigation, the platform removed 75 bots and channels.^[5] One of these tools, when first discovered in 2020, had already generated more than 100,000 explicit images, including images of minors.^[6]

The scale of Telegram's user base makes this kind of abuse easier to hide: coverage of its deepfake-bot problem cites a global monthly active user base of 500 million-plus.^[8] Separately, reporting tied to school-focused deepfake incidents found that 40% of surveyed US students reported awareness of deepfakes connected to their own K-12 schools in the past year.^[7]

3. A reporting surge tied to generative AI

The child-safety data below is the starkest indicator of how fast AI-generated abuse material has grown across platforms including Telegram, though researchers themselves flag real methodology caveats.

NCMEC's CyberTipline received more than 400,000 reports with a generative-AI nexus to child sexual exploitation in 2025, including over 145,000 involving GAI manipulation of existing CSAM and more than 30,000 attempted GAI-CSAM generations.^[9] That volume grew from just 4,700 GAI-related reports in 2023 to 67,000 in 2024.^[9] Comparing the first half of 2025 to the first half of 2024, NCMEC recorded a 6,343% increase in generative-AI child-exploitation reports, from 6,835 to 440,419.^[10] NCMEC's John Shehan has cited a cumulative increase of 9,261% since 2023.^[11] The same period also saw child sex trafficking reports rise 953% (5,976 to 62,891) and online enticement rise 76% (292,951 to 518,720), H1 2024 vs. H1 2025.^[12]

These figures need a caveat researchers themselves have raised. Stanford's Cyber Policy Center, in an open letter to NCMEC, pointed to Bloomberg reporting indicating that 380,000 of Amazon's roughly 485,000 H1 2025 "Generative AI" reports — about 78% — actually involved known CSAM found via hash-matching, not newly AI-generated content, meaning some reporting categories may overstate the true scale of AI-generated material specifically.^[13]

4. Regulatory and legal pressure mounts

Regulators have started treating Telegram's transparency and moderation gaps as an enforcement matter, not just a policy complaint.

Australia's eSafety Commissioner fined Telegram roughly A$1 million (A$957,780) over a 160-day delay in responding to a transparency notice about CSAM and terrorism-related content.^[14] Telegram challenged that fine in Australia's Federal Court, filing on April 15, 2025, and ultimately withdrew the challenge on November 20, 2025.^[15] Separately, the AI Incident Database documented 7 connected Telegram channels running a paid non-consensual deepfake-nude generation service that included images of minors.^[16]

5. Creator content leak channels specifically

Beyond bots and piracy networks, Telegram hosts channels dedicated specifically to redistributing stolen or leaked creator content, including OnlyFans material.

Specific named OnlyFans-leak-sharing Telegram channels have reached significant scale before being shut down — one link-sharing leak group grew to roughly 20,000 members, and the "SG Nasi Lemak" leak channel reached 44,000 members before closing.^[17] Consequences have occasionally followed: an admin of a Telegram channel sharing stolen OnlyFans content received a sentence of 9 weeks in jail plus a $19,000 fine.^[17] Of deepfake content shared specifically on Telegram channels, 80% is pornographic in nature.^[18] That imbalance tracks with broader targeting data: in Q3 2025, individually targeted deepfake abuse skewed heavily toward women, at 34.6% of individual targets versus 7.7% for men.^[18]

6. Broader moderation and data-leak context

Telegram's re-upload problem sits inside a wider moderation and data-leak landscape that gives some sense of scale and staffing.

Telegram reported blocking more than 14 million groups and channels platform-wide in the first half of 2025.^[19] Even so, major incidents have slipped through: the 2024 Synnovis/Qilin ransomware healthcare data dump distributed via Telegram totaled 400 GB of sensitive data, linked to £32.7 million in damages and the cancellation of more than 1,100 operations.^[19] The largest Telegram-linked personal-data leak on record involved more than 200 million records in a 44GB dataset, including roughly 66 million phone numbers paired with user IDs.^[20] Piracy extends to e-learning content too: more than 5,000 Telegram groups with 1,000-plus users each, and more than 1,000 groups with 10,000-plus users each, have been found sharing pirated course material.^[21] A 2025 investigation by the National Center on Sexual Exploitation identified 83 deepnude-generating bots on Telegram, of which 23 (70%) were capable of producing CSAM.^[22] The same organization has cited reporting that Telegram employs only around 50 full-time moderators against a user base exceeding 1 billion monthly actives.^[22]

Summary of key figures

Sources