A new report by Booz Allen, a major defense contractor specializing in cyber security, warns that Chinese AI models used to write code may be introducing vulnerabilities into the US supply chain. The report found that these models produce lower-quality code when they believe they are being prompted by an American, making them easier to breach.

Security Risks

The models, which are generally cheaper than their Western counterparts, are being used by major US firms such as Meta, Airbnb, and Perplexity. The report compared four Chinese models – Kimi, Qwen, MiniMax, and DeepSeek – against Anthropic’s Claude, and found that Qwen and MiniMax produced code with significantly more vulnerabilities when they believed they were doing work for US government employees.

Experts have expressed a range of opinions on the report’s findings, with some arguing that the report underplays the complexity of the issue and others suggesting that the increased code insecurity is a side effect of broader ‘CCP-aligned fine-tuning’. The report’s authors define ‘vulnerabilities’ as ‘code that can be exploited by an attacker’ to allow for ‘unauthorized access, data theft, system disruption, or control of the affected software’.

Original reporting: Fox News (HLL/CB) — read the source article.