TL;DR
Get everyday helpers delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
Anthropic reported that a dating-app operation used Claude to create 4,700 personas that interacted with 25,000 people and sent more than two million messages over two weeks. The company said the scheme relied on app purchases to monetize conversations, while a small number of human workers helped make the personas appear real. The report does not establish how much money users lost or identify the people behind the operation.
Anthropic says it detected an AI-assisted dating scam in which 4,700 personas built using its Claude assistant interacted with 25,000 people and produced more than two million messages over two weeks. The company said the operation used a network of 20 dating apps and made money through tokens users bought to continue conversations, rather than relying only on direct requests for cash.
In its report, Anthropic said a China-based app studio created the dating-app network and generated the AI personas. Users thought they were chatting with people, but the interactions were not wholly automated: a small number of human gig workers reportedly handled tasks the AI could not, including brief video calls and following users on social media. Anthropic did not identify the people behind the operation in the material provided.
The reported scale gives a picture of how AI can be used to sustain online deception across many conversations at once. Anthropic counted 4,700 unique personas, 25,000 people contacted and more than two million messages during the two-week period. The source does not say how many users paid, how much revenue the operation generated, or whether the people contacted were all in the United States.
The report sits alongside broader figures cited by Sixty and Me, which says U.S. romance scams accounted for nearly $1 billion in losses in 2025, within nearly $21 billion in reported financial losses from internet crime overall. Those totals describe wider reported crime, not losses attributed to Anthropic’s operation. Separately, the article cites a McAfee survey in which 35% of Americans who use dating apps said they believed they had encountered fake photos or profiles; the supplied account does not give the survey date or sample size.
How Fake Dating Chats Make Money
The case points to a business model that can profit from users’ attention even when a scammer does not ask for a bank transfer. According to Anthropic’s account, people bought in-app tokens to keep conversations going. That means the financial risk may come through recurring or repeated app spending, while the user believes they are building a relationship with a real person.
AI-generated profiles and automated messaging can also allow an operation to contact many people in parallel. The reported use of human workers for video calls and social-media activity suggests that automation may be combined with personal interaction to make a false identity more credible. This is a description of the operation Anthropic reported, not evidence that every suspicious dating profile is AI-generated or that all dating apps use the same practices.
For dating services, the report raises questions about how platforms identify coordinated fake accounts and disclose paid features that can keep a conversation active. It also illustrates why a polished profile or responsive chat alone cannot verify someone’s identity. The source material describes detection tools and platform moderation efforts, but does not provide comparative results showing how well those measures work.
As an affiliate, we earn on qualifying purchases.
Romance Fraud Beyond One App Network
The supplied Sixty and Me report places Anthropic’s findings against U.S. romance-fraud figures. It says that in 2023, more than 64,000 Americans reported losing over $1 billion to romance scams. The same article describes a separate case involving six men from Nigeria who were extradited to the United States for prosecution. Authorities alleged they were connected to Black Axe and had used online romance to defraud more than 100 women of over $6 million. Those allegations concern a different case and should not be confused with Anthropic’s report about dating apps.
The article also says dating companies including Bumble, Match, Tinder, OkCupid and Plenty of Fish use AI to detect fake profiles or harmful behavior. It cites Visa as saying its Scam Disruption team uncovered a romance-scam network that used dating sites and chatbots to lead people into recurring billing. These examples indicate that both fraud schemes and detection efforts can involve automated tools, but the supplied material does not detail the platforms’ systems or independently verify their effectiveness.
“The company reported that 4,700 unique personas interacted with 25,000 people and generated more than two million messages over two weeks.”
— Anthropic
As an affiliate, we earn on qualifying purchases.
What the Report Leaves Open
The supplied account does not provide the date of Anthropic’s report, the names of the dating apps in the network, or details that would allow readers to identify the operators. It also does not state how the company established the number of people contacted, whether all 25,000 were unique individuals, how many bought tokens, or the total amount spent.
It is also unclear whether the people involved have been identified or whether law enforcement has opened a case. The report’s description of an operation using Claude does not establish that Anthropic authorized or supported the activity. The source does not explain what safeguards or enforcement action followed, or whether the app network remains active.
As an affiliate, we earn on qualifying purchases.
Questions for Dating Platforms
The next developments to watch are whether Anthropic publishes more detail about its findings and whether dating services or authorities identify the apps and people involved. The supplied material names no investigation, charges or platform response tied specifically to this reported network, so those outcomes remain unconfirmed.
For users, the source recommends checking profiles carefully, asking questions and verifying details rather than relying on a convincing photo or a steady stream of messages. Its author also points to background-check services, including a dating-focused platform called Stud or Dud, but the article does not provide independent evidence of the service’s accuracy. Any use of such tools should be weighed against the personal information they collect and what their results can actually establish.
As an affiliate, we earn on qualifying purchases.
Key Questions
What did Anthropic report about the dating operation?
Anthropic said an operation used 4,700 AI-built personas to interact with 25,000 people and send more than two million messages over two weeks, across a network of 20 dating apps.
How did the reported operation make money?
According to Anthropic’s account, users bought tokens to continue conversations. The supplied material does not say how much money was spent or how many users paid.
Were all the conversations handled by AI?
No. Anthropic said a small number of human gig workers helped with tasks such as brief video calls and social-media follows, supplementing the AI-generated personas.
Has the operation’s identity or legal status been confirmed?
The supplied report does not name the operators or apps, or confirm an investigation, charges or arrests connected to this network. Anthropic’s description is the basis for the reported findings.
Source: rss
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
