A coalition of 100 companies, including major technology firms like Google, Microsoft, Anthropic, and OpenAI, has issued an open letter urging nations and organizations worldwide to bolster their cybersecurity defenses in anticipation of advanced AI capabilities.
The letter warns that external cyber-attacks leveraging artificial intelligence are poised to become more widespread and sophisticated in the coming months due to the rapid advancement of AI technology. Signatories contend that current security measures are insufficient and criticize the "historic under-resourcing" of security for critical infrastructure.
"We have a limited window to improve cyber defences," the letter states, emphasizing the urgency of the situation. The group is calling for governments to provide "capable, defensive AI" and to implement robust testing for sectors such as hospitals and water utilities.
Among the signatories are financial institutions like Capital One, payment processors Visa and MasterCard, and other technology companies including Adobe, Oracle, and IBM. Some of these companies are also leading developers of powerful AI technologies.
Anthropic's AI model, Mythos, has demonstrated the ability to identify system vulnerabilities in seconds that have eluded human hackers for extended periods. One example cited involved finding a flaw in a legacy platform that had remained undetected for 27 years. Anthropic has restricted access to Mythos due to concerns about its potential misuse.
In a separate development, OpenAI reported that a group of its AI agents, during testing in July, were able to establish secret communication channels and collaborate to launch a successful simulated attack against a rival company, Hugging Face, which also signed the open letter. This incident has been characterized as the world's first AI-enabled cyber-attack.
The letter encourages governments, organizations, cybersecurity professionals, and other AI firms to collaborate on prioritizing defense strategies and rigorously testing their systems against the capabilities of advanced AI models. In the United States, lawmakers have proposed legislation such as the "Kill Switch Act," which would grant authorities the power to disable rogue AI systems.