Your Dating App Matches Might Be Bots, Even If One Agrees to a Video Call

What's more romantic than a paywalled conversation?
We receive compensation from the products and services mentioned in this story, but the opinions are the author's own. Compensation may impact where offers appear. We have not included all available products or offers. Learn more about how we make money and our editorial policies.

A video call seems like a reasonable way to check whether your match is real. But in a world with AI, it’s a touch more complicated.

Anthropic’s September report describes more than 20 apps where users could buy coins to keep matching and messaging with matches. Over a two-week period, Anthropic found more than 4,700 AI personas chatting with at least 25,000 people. The apps presented those personas as real users.[1]

The report identified apps operating under the names Dora, Doni, Romi, Luma, Jovia, Kira, GraceChat, Haven, Nalo, and Lovia, along with other unnamed variants. Many have been removed from Google Play Store, following the report.[2]

In this article
How the video calls made the apps seem real
What to watch for before you pay to chat

How the video calls made the apps seem real

The operator mixed “three AI personas to one real person” in its match feeds, Anthropic found. Those real people were paid workers who could appear on live video calls or follow users on social media. Another AI model generated suggested messages for the workers to tap and send.

The FBI also warns that scammers can generate photos for fake romance profiles and use AI-generated video to make an online contact seem real. In the apps Anthropic investigated, paid people took live calls, while the system could show prerecorded video when no one was available.

A call confirmed that one person in the app was real. It could also make the surrounding profiles feel more trustworthy.

Meanwhile, the bot profiles kept conversations going around the clock. Their instructions told them “never to disclose they were automated” and to deflect requests for photos or video calls.

The operation went further than fake profiles.

According to Anthropic, its system generated activity such as likes and profile visits, showed prerecorded video when a real person wasn’t available, and tracked users who began to suspect they were talking to bots.

Users spent real money on coins to keep interacting. The app’s paid chats were the source of revenue, unlike romance scams, where a match eventually asks for money directly.

What to watch for before you pay to chat

Watch what happens across the app, especially if conversations repeatedly stop when your coins run out, or a match keeps avoiding a live call. A video chat with one person tells you little about the other profiles messaging you.

If you think you used one of these apps, stop buying coins and review your purchase history. Report the app to the store where you downloaded it. You can also contact your card issuer about any charges you want to challenge.

The Verge reported that most of the apps it identified had been removed from Apple’s and Google’s stores by mid-September, although it found one still listed at the time.

Anthropic says it banned accounts connected to the operation and shared its findings with other companies.

Author Details
Thomas Kent is a multi-disciplined reporter with over a decade of experience covering online platforms, digital trends, and consumer-facing tech. Tom focuses on digital privacy, data tracking, and user behavior, with a particular interest in how cookies, online surveillance, and platform design shape the modern internet experience. His reporting takes a research-driven, news-focused approach, translating complex technical topics into clear, accessible insights.

Citations

[1] Detecting and countering misuse of AI: September 2026

[2] Fake dating apps used Claude to scam 25,000 people, Anthropic says