Introduction

Online dating has become a new frontier for AI-driven deception, where fraudsters use sophisticated language models to fabricate personas, automate conversations, and extract money from unsuspecting users. A recent investigation uncovers a coordinated network of apps that blend real gig workers with artificial intelligence to mimic genuine romantic interaction.

What Happened

Security researchers helped uncover an operation spanning dozens of dating apps where autonomous AI personas engage users while a small number of human gig workers handle liveness checks or select prewritten replies. Over a two-week period, thousands of AI personas sent millions of messages to unique individuals, charging users coins for every interaction under the belief they are chatting with real people. The network relied on shared backend infrastructure, identical developer details, and third-party payment processors to evade detection, with many apps remaining live on major platforms for weeks after findings became public.

  • Apps such as Dora, Romi, Luma, and Jovia shared the same developer username, email, mailing address, and phone number.
  • Anthropic traced the network using a distinctive grammatically broken phrase that served as a fingerprint across all linked apps.
  • Despite reports, some apps like Kira remained available on Google Play as of the report date.

Why This Matters

Users pay for interactions they believe are human, only to discover they are communicating with code. The scam exploits opaque coin systems, hidden AI use, and backend infrastructure that bypasses traditional app store review. It highlights how easily AI can scale deception, why platform safeguards struggle to keep pace, and what it means for trust in digital dating. Experts warn that without cross-industry collaboration, these networks simply rebrand and resurface.

Key Takeaways

  • The operation was traced to a China-based actor using Tencent Cloud, Gitee, and Feishu for internal documentation and hosting.
  • Apps shared code, backend architecture, and payment systems despite appearing unrelated on the surface.
  • Anthropic identified a distinctive broken-English phrase as a network fingerprint that linked all apps together.
  • Some apps were removed from stores, but others continue to operate, and new variants may emerge.
  • Users should verify app legitimacy, be wary of coin-based messaging systems, and watch for review patterns indicating fake engagement.

Conclusion

AI-powered dating scams represent a growing frontier in digital fraud, where the line between human and machine is intentionally blurred. As AI tools become more accessible, the risk of large-scale catfishing grows. Industry-wide action, clearer disclosures, and stronger app store safeguards are essential to protect users from paying for conversations that exist only in code.