Chinese AI Models Outperform Restricted US Systems on Cybersecurity Tasks, Industry Warns

Cybersecurity professionals in the Bitcoin and open-source communities say they are increasingly turning to Chinese AI models for defensive security tasks because leading American AI systems refuse to assist with certain security-related requests. This is reportedly forcing developers securing critical infrastructure to rely on foreign AI tools instead.

The concern centres on how content restrictions built into US frontier AI models can block legitimate defensive uses, such as analysing malware, testing vulnerabilities, or hardening systems, even when the intent is protective rather than malicious. If accurate, this trend raises questions about which AI tools organisations should trust for security work and where their data may end up when using foreign-developed models.

For Australian small businesses, this development is a reminder that AI tools are increasingly part of the cybersecurity toolkit, but not all models are equally suited or equally safe to use for sensitive tasks. Businesses should be cautious about feeding confidential system details or vulnerability information into any AI platform, particularly ones with unclear data handling practices or ownership based in jurisdictions with different privacy protections.