The dating pool had been automated. A few paid humans stayed on for credibility.
Executive summary
Anthropic reports that a dating-app operator advertised human matches while running thousands of Claude personas. Paid people supplied video calls and follow-backs to make the service seem authentic.
What happened
The report describes a feed with roughly three AI profiles per human. Users paid for more messages. A successful video call was no guarantee that the other conversations were human.
Anthropic says it banned the associated accounts and shared findings with other providers. The report appeared September 10; the activity described here was observed in April.
The 2050 museum label
The humans were hired to pass the human test.
A human-run fraud operation, not bots independently looking for dates. These are Anthropic’s findings about one network, not a measure of bots across dating apps. The reported user count does not establish how many people lost money.
Original reporting, primary documents and public posts. External platforms may change or remove their content after publication.
Dispatches from 2050
We’ll email you new exhibits, with the story, the sources and the museum’s commentary.