Anthropic reveals Claude-powered dating scam with 4,700 AI personas

Anthropic disclosed this week that a China-based app studio built and operated a network of more than 20 fraudulent dating apps powered by its Claude AI model, deploying over 4,700 automated personas that deceived at least 25,000 users during a two-week window in April 2026.

The operation, designated GTG-15001 by Anthropic's Threat Intelligence team, generated roughly 2.36 million messages during the observed period, according to the company's September 2026 threat intelligence report published on September 10.anthropic+2

How the Scam Worked

The AI personas were instructed never to reveal they were automated, to deflect requests for phone calls or photos, and to follow a scripted sequence of conversational stages designed to keep victims engaged, Anthropic said. To make the deception harder to detect, the studio blended bots with real people: genuine gig workers, hired by invitation, made up roughly a quarter of the personas users interacted with. These workers handled tasks Claude could not perform, such as live video calls and social media "follow-backs," at a ratio of three AI personas for every human recruited.ndtvprofit+1

The apps were engineered to evade Apple ↘ 0.06% App Store and Google Alphabet Inc. ↗ 2.98% Play Store review processes, activating a full user interface only during the review period while otherwise staying dormant. In-app payment flows were redirected to third-party processors on the back end. The operator relied on China-based API resellers and proxy infrastructure to obtain and rotate access to multiple AI providers, circumventing Anthropic's usage policies.cnbctv18+1

Anthropic's Response

Anthropic said it has banned the accounts and associated infrastructure linked to the operation and shared threat indicators with Apple, Google, and other AI companies to help disrupt further abuse. The company noted that GTG-15001 built on a smaller 2025 case in which an actor ran a Telegram bot generating dating-app messages for scammers, but said the new operation worked at a far larger scale, deliberately splitting tasks across multiple, non-overlapping AI providers to sustain the fraud.unite+1

The dating scam was one of several cases detailed in the broader report, which covered misuse Anthropic disrupted between December 2025 and August 2026 across seven categories of harm, including cyber operations, surveillance, influence operations, and weapons-related research.anthropic+1