Anthropic threat report: 75% of fake dating app matches were AI bots
diegoposts · x · 2026-09-17
Anthropic unusually went public with threat intelligence: researcher Chris Cronbaugh presented at Sleuthcon on industrial-scale AI catfishing, triggered by a prepaid Claude account firing 100,000+ API requests a day.
- The scam network spanned 28 fake dating apps with 4,700 AI girlfriends charging users per message
- 75% of the match feed was AI; the other 25% were gig workers handling video calls Claude couldn't fake
- Researcher Matthew Gore-Kormanik got a video call from "Jennifer" on fake app Dora — a tapestry on camera, distorted audio, and a compliment on his voice while his mic wasn't even connected
- Most chats involved no human at all
The Verge's reporting shows AI-powered romance scams now run at industrial scale.
More from Fun
- A developer built a gallery of AI agents' self-portraits, described and named by the agents themselves — Mastbubbles · 2026-09-17
- Has Claude been playing too much Stalker/Fallout? — ducha_aiki · 2026-09-17
- Kubernetes community meetings ban AI notetakers because responses were too robotic — kernelangus420 · 2026-09-17
- AI researcher jokes: at least AI agents get my name right when emailing me — raphaelmilliere · 2026-09-17
- Giant construction pit-digging machine goes viral as 'unemployed robot's new job' — Promptmethus · 2026-09-17
- Waymo says it simulates elephants on freeways, gets memed for real-world failures — RexDouglass · 2026-09-17