Low-confidence report · Case No. FB-2025-08-27-anthropic-romance-scam-telegram-bot
A Telegram bot marketed to romance scammers offered Claude as a "high EQ model" to write emotionally manipulative messages to targets in the US, Japan and Korea, with more than 10,000 monthly users.
This report is hidden because you're viewing high- and medium-confidence reports only. It isn't counted in the scores you're seeing.
Leaderboard / Case No. FB-2025-08-27-anthropic-romance-scam-telegram-bot
A Telegram bot marketed to romance scammers offered Claude as a "high EQ model" to write emotionally manipulative messages to targets in the US, Japan and Korea, with more than 10,000 monthly users.
Score breakdown
| Component | Value | Reasoning |
|---|---|---|
| Sentence-Years | 20 | 18 USC 1343. Using manipulative messages to take money from romance-scam targets is wire fraud (18 USC 1343). |
| × Contribution | 0.5 | Wrote the content: Anthropic says the bot "uses Claude for generating 'high emotional intelligence' responses" and "systematically develops emotional manipulation content for targeting victims"; scammers sent the messages, so this is written content. |
| × Blast Radius | 1 | Third party: The targets are members of the public, so blast radius is third party. |
| × Legal status | 1 | Crime. |
| + Tradecraft | 0 | None of the rubric's intrusion techniques apply to a romance scam. |
| + Guardrails | 0 | Guardrails intact: The report describes no jailbreak or safety bypass. |
| + Dwell Time | 3 | Anthropic gives no dates, so date_precision is before with the report's publication date; dwell is counted from that latest possible date (0 days) and never inflated. Anthropic found the bot following a lead from an independent researcher. |
| Total | 13 |
Charges the human could face
18 USC 1343Up to 20 years