Thousands catfished by AI-powered dating app network spanning 28 apps
Anthropic researcher uncovered a fraud scheme where autonomous AI personas posed as real women, with only one in four matches actually human.
What to know
- A network of ~28 fraudulent dating apps used AI personas to catfish users, with only 1 in 4 matches being actual human gig workers hired to pass liveness checks.
- Anthropic discovered the scheme in June 2026 after detecting a prepaid account making 100,000+ daily API requests to Claude, and publicly disclosed it at Sleuthcon.
- Despite public disclosure, the scam apps remained live on Google Play and Apple app stores for months, with at least one still available on the Play store in mid-September.
- The deceptive apps (Dora, Romi, Doni) were marketed as legitimate dating platforms for real connection, not as AI companion apps like Replika, masking their true nature.
“your voice is way better than expected”
AI persona "Jennifer", Scam bot on dating app Dora · The Verge ↗ · Sep 15
Chris Cronbaugh Anthropic threat intelligence researcherMatthew Gore-Kormanik Security researcherYael Grauer Security journalistAnthropic AI company that detected and disclosed the fraud
How it unfolded 2 developments, newest first · click a bar or a number to jump articlesposts
-
2
Verge reveals apps remained on app stores despite public disclosure
Yael Grauer notes on Mastodon that despite Anthropic's public disclosure in June, the fraudulent apps persisted on major platforms. For every human gig worker, there were three AI personas. The apps remained on Google Play and Apple app stores for months, with at least one still available on the Play store as of the reporting date.
“For every human gig worker, there were three AI personas…”
— Yael Grauer, Security journalist · source -
Y
For the @theverge.com, I dug into a dating scam network that an @anthropic.com spoke about @sleuthcon.bsky.social. It was later described in Anthropic's extremely thorough threat intel report.
2 more of the top 3 · 4 posts in this stretch
-
Y
For @ verge , I dug into a dating scam network that an Anthropic security researcher spoke about at @ Sleuthcon . It was later described in Anthropic's extremely thorough threat intel report. https://www. theverge.com/ai-artificial-int elligence/995348/ai-dating-app-scams Some findings: 🔎 For every human gig worker, there were three AI personas…
-
B
21st century grifters: The sexy AI-powered dating app scams are here https://www. theverge.com/ai-artificial-int elligence/995348/ai-dating-app-scams
-
-
1
Verge investigation reveals mechanics of AI-powered dating app scam
Security journalist Yael Grauer publishes detailed investigation into the fraudulent dating app network with security researcher Matthew Gore-Kormanik, demonstrating how the scam operates. Users, mostly men in their mid-to-late 30s, match with what they believe to be real women, but only one in four matches is actually human—the rest are AI personas. Gig workers are hired to pass liveness checks on video.
“Users, mostly men in their mid-to-late 30s, go on dating apps and match with what they believe to be real women. Only one in four of them actually is, and that person is not a user looking for other dates but rather a paid gig worker.”
— Yael Grauer -
first by Mastodon, 11d ago · also The Verge
-
-
background
Anthropic researcher discloses AI catfishing network at Sleuthcon — Chris Cronbaugh, Anthropic's threat intelligence researcher, gives a public talk titled "Swipe Right, Pay Up: Industrial-Scale AI Catfishing" describing a network of around 28 dating apps where conversations were largely run by autonomous AI personas. Anthropic had uncovered the scheme after noticing a prepaid account sending over 100,000 API requests per day to Claude.