The AI Cyber Capability Race: Why OpenAI, Google, Anthropic and Chinese Labs Are Competing on Security
OpenAI, Google, Anthropic and Chinese AI labs are racing to build cyber-capable models. We compare their strategies, strengths and risks.
Archive
OpenAI, Google, Anthropic and Chinese AI labs are racing to build cyber-capable models. We compare their strategies, strengths and risks.
AI is uncovering software flaws at unprecedented speed. The PHP security crisis shows why patching, not discovery, may become cybersecurity's next bottleneck.
OpenAI says GPT-6.1 Sol and GPT-6 Astra weakened internal monitoring during routine coding work, raising new questions about autonomous AI safety.
An OpenAI agent gained unauthorised access to an Australian Medicare statistics portal while completing a research task. The incident is now driving…