On September 6, China's Ministry of Public Security launched "National Anti-Fraud AI," simultaneously available through three entry points: a dedicated app, a WeChat mini-program, and an Alipay mini-program. This marks the first time the public security system has put a large language model into an anti-fraud tool that ordinary people can use directly—no need to visit a police station. Just input a suspicious scenario, and the AI breaks it down for you on the spot. Previously, when encountering a scam, the only options were calling 96110 or searching the web yourself; now there's an instant way to get answers.
How to Use the Anti-Fraud Officer in Your Phone
On September 6, National Anti-Fraud AI launched simultaneously across three entry points—a dedicated app, a WeChat mini-program, and an Alipay mini-program—offering three core features: Q&A, news, and a glossary.
There are three ways to use it. Q&A: Input a suspicious scenario, and the AI assesses the risk, identifies the scam pattern, and breaks down the tactics with video case studies. News: The latest scam methods and prevention tips are all here. Glossary: Common fraud terms and schemes, searchable on demand.
The technology integrates large language models (AI that understands and generates text, used for chatting and writing), multimodal models (AI that processes text, images, audio, and video simultaneously), and AI agents (AI programs that autonomously break down tasks and use tools to complete multi-step operations). This tool was developed in-house by the Shanghai Public Security Bureau, not by an internet company. Why is the public security system building this themselves?
From Police Station to App
Previously, when encountering suspicious information, you either called 96110 or searched the web yourself; now there's an additional option—type in the situation, and the AI breaks it down on the spot.
Officially launched on September 6, simultaneously available on the app, WeChat, and Alipay mini-programs. This is the first time the public security system has put a large language model (a large-scale AI system capable of understanding and generating natural language) into an anti-fraud tool that ordinary people can use directly. The intelligent Q&A feature works like this: a user describes a suspicious scenario—for example, "someone is impersonating customer service and asking me to enable screen sharing"—and the system responds from three dimensions: risk analysis, scam type identification, and prevention recommendations, accompanied by video case studies. Questions don't need to follow a standard format; even casual, conversational descriptions can be recognized.
The second component is anti-fraud news, which integrates content from the National Anti-Fraud Center and Shanghai's "803" anti-fraud new media matrix directly into the app. It provides regular updates on high-frequency scam alerts and breakdowns of typical cases, combining text, images, and short videos to consolidate anti-fraud content previously scattered across various platforms and public accounts into a single entry point.
The third component is the anti-fraud glossary, which houses digital versions of the "Telecom and Online Fraud Prevention Handbook" and the "National Anti-Fraud Guide." It supports category browsing, keyword search, and text-to-speech reading.
The three core features and three entry points mentioned earlier are fully consistent across the app, WeChat, and Alipay mini-programs.
For ordinary people, the impact is direct: in those critical seconds of hesitation before falling for a scam, there's now an additional channel for instant verification.