The Ministry of Public Security has highlighted a rise in criminal activities leveraging artificial intelligence, prompting a multi-pronged crackdown as China embarks on its "15th Five-Year Plan" period.

During a press conference on July 24, senior officials outlined measures to enhance governance, protect citizens, and foster development. The ministry's deputy head,亓延军, noted that China's ability to manage risks from technologies like AI and low-altitude security has markedly improved, alongside a more robust system for cyberspace security governance.

In recent years, criminals have used AI for various illegal acts, including telecom fraud, privacy breaches, and spreading disinformation that undermines social trust and national security. In response, authorities have intensified efforts to identify and dismantle these operations.

Key areas of AI-driven crime

Liu Zhongyi, head of the ministry's Criminal Investigation Bureau, detailed three primary manifestations of these new crimes. First, criminals use AI to forge identities and infiltrate government or corporate systems to steal information and assets. Second, they impersonate acquaintances or authoritative figures to gain trust and commit fraud. Third, AI tools are employed to generate and circulate online rumors, disrupting public order.

Authorities are not only targeting the perpetrators but also cracking down on criminal networks that provide technical support for these activities, aiming to eradicate the root cause of such offenses.

Enforcement results and preventive measures

In the first half of this year, police investigated over 170 cases involving AI-generated online rumors, leading to legal action against more than 190 individuals.

To further counter these threats, the ministry is leveraging big data analytics to build risk assessment models. These models analyze suspicious activities such as frequent unknown video calls, silent audio recording requests, multiple rapid facial recognition attempts, and video calls initiated from overseas. This analysis generates precise early-warning alerts, which are then sent to local police departments for proactive intervention and warnings to potential victims.

Additionally, authorities are monitoring security risks within major AI models themselves. They have conducted over 600 inspections of AI application services this year, issuing warnings, ordering rectifications, or shutting down more than 30 service providers for non-compliance.