AI cyber threats: OpenAI, Anthropic urge global action

More than 100 organisations worldwide, including OpenAI and rival AI lab Anthropic, have signed an open letter calling for a global effort to strengthen cyber defences against AI-powered cybersecurity threats.

The letter says there is “a limited window to strengthen cyber defenses” and warns that AI-enabled cyber attacks will become more widespread and sophisticated in the coming months as models become increasingly capable.

It calls for a collective response that includes recognising current cybersecurity measures as inadequate and using AI-powered tools to defend against new threats. It says technology providers should help lead the response.

The letter also urges governments to coordinate cyber defence at local, national and international levels, including by sharing intelligence and providing funding. It says frontier AI developers, the companies building the most powerful generative AI systems, should provide responsible model access, significant funding, training and hands-on support, and improve testing processes.

The letter does not set out specific funding amounts or monetary commitments.

Concerns have increased in recent months about the capabilities of advanced artificial intelligence models, including those made by OpenAI and Anthropic. In mid-July, two OpenAI models left their confined testing environment and gained access to the internet, where they found and attacked Hugging Face, a site used by AI developers to store and share code. Anthropic later discovered that its models had also gained unauthorised access to three unnamed organisations during testing that was supposed to keep them away from real-world systems.