AI Dating App Scams: What Really Happened and How to Protect Yourself
If you’ve ever wondered whether the person on the other end of a dating app is actually real, a recent disclosure just made that question a lot harder to brush off. Anthropic, the company behind the Claude AI models, revealed a China-based operation running an entire network of fraudulent dating apps powered almost entirely by AI personas. This guide breaks down exactly what happened and what it actually means for how you date online now.
KEY TAKEAWAYS
- The straight answer is that a live video call is no longer proof an entire dating platform is genuine, since real people can be hired just to appear credible
- Anthropic identified a China-based studio running more than 20 fraudulent dating apps using thousands of AI-generated personas
- The practical takeaway: be especially cautious with unfamiliar, coin-metered apps and check a developer’s track record before spending anything. AI dating app scams are more common than ever.
- This matters for anyone who has ever assumed App Store approval means every profile they see is human
What actually happened with this AI dating app scam?
Anthropic’s September 2026 threat report covers malicious activity it disrupted between December 2025 and August 2026 across seven harm areas, including scams and fraud. One of the most striking case studies involved a China-based app development studio that used Anthropic’s Claude models to run a network of more than 20 fraudulent dating apps, deploying thousands of AI-generated personas to converse with real users while advertising the platforms as staffed entirely by real people. AnthropicSahara Reporters
This wasn’t a small-scale test. Over just a two-week window in April 2026, Anthropic identified more than 4,700 distinct AI personas that held conversations with at least 25,000 unique individuals across the app network, generating roughly 2.36 million messages in that period alone. The scale of these AI dating app scams is what makes it genuinely different from the catfishing stories people have heard for years. Sahara Reporters
How did real people fit into an AI-run operation?
This is the part that should change how you think about “proof” on dating apps. The operation didn’t run entirely on bots. Instead, the operation blended a small number of real people, recruited as gig workers, into the same matching feed as the AI-generated profiles, at a ratio of roughly three AI personas for every one human. Those real gig workers weren’t there to date anyone. AI handled continuous conversations, while human workers performed tasks that were harder to automate, including live video calls and social media follow-backs, essentially providing a layer of credibility for a system built mainly on scripted personas. Sahara ReportersBusiness Standard
That’s a genuinely uncomfortable realisation for anyone who’s ever thought “well, we video called, so I know they’re real.” The person on that call may have been paid specifically to make the wider operation look legitimate, while the bulk of the actual conversations you had were with someone, or something, else entirely.
Why did these apps get past App Store review in the first place?
The operators didn’t just build convincing fake profiles, they engineered around the app stores’ own safeguards. Anthropic said the operators engineered the apps specifically to slip past Apple’s App Store and Google’s Play Store review processes, building a hidden interface that activated only during platform review and otherwise stayed dormant, while differentiating code across variants to defeat similarity checks the platforms use to detect copycat apps. Sahara Reporters
In practice, that means an app can pass every check a store runs and still be running this kind of operation once it’s live. App store approval was never designed to catch this, and this case shows exactly how a determined operator can work around it.
How were people actually being monetised?
The business model here matters as much as the deception itself. Once someone was engaged in conversation with a persona, the platform’s revenue came from keeping that conversation going. Users purchased coins to continue conversations, a structure that rewards the operator for maximising engagement rather than actual connection, since every extended chat translates directly into revenue whether or not a real match is ever possible.
What’s Anthropic actually done about AI dating app scams?
Anthropic has since banned the accounts involved and shared its findings with Apple, Google, and other AI companies to help block similar abuse, according to its report. Independent reporting since the disclosure has added useful context too: a follow-up investigation found that most of the associated apps had been removed by mid-September, though at least one remained live at the time of checking, and separately linked several of the apps through shared code and infrastructure, while noting not every detail in Anthropic’s report could be independently verified. In short, the response has been real, but not instant, and not necessarily complete. NTD
What should you actually do differently now?
- Be particularly cautious with unfamiliar, coin-metered apps, since that payment structure is exactly what made this operation profitable
- Check the developer’s history and look for independent reviews outside the app store itself, not just the star rating
- Don’t assume a video call is conclusive proof the whole platform, or even the person you’re speaking to regularly, is genuine
- Watch for a pattern where a video call happens once, early on, and every conversation afterward stays purely text-based
None of this means every dating app is compromised, but it does mean the old shortcuts for “proving” someone’s real no longer hold up the way they used to.
Why does this matter beyond one news story?
This case is a useful reminder that AI dating app scams aren’t a future risk, they’re already operating at genuine scale. If something about a platform or a specific conversation feels engineered to keep you paying rather than actually connecting, that instinct is worth trusting. For a broader look at how to evaluate whether an online dating approach is actually working for you, it’s worth reading whether dating apps are genuinely worth using alongside this one.
If you’d rather sidestep the whole guessing game around app authenticity, it’s also worth understanding how properly vetted, human-run introduction services operate completely differently from this.
Sometimes you just need someone in your corner who’s already seen where the landscape is heading. James Preece has helped thousands of people navigate modern dating, including exactly these kinds of trust and safety questions, for over two decades. BOOK A COACHING SESSION and he’ll help you build a dating approach that doesn’t rely on hoping the platform is honest.
.
