OpenAI, Google, and others urge stronger cybersecurity for critical infrastructure

A coalition of leading AI and cybersecurity companies, including OpenAI and Google, has issued a stark warning: the window to prepare for increasingly sophisticated AI-enabled cyberattacks is closing rapidly. In a joint letter, the companies emphasized that the "status quo security won't be enough" and called for coordinated action and increased funding for cybersecurity programs. They specifically highlighted the need to bolster protections for critical infrastructure, encompassing essential services such as hospitals, water treatment facilities, and local authorities. The signatories propose expanding threat intelligence sharing and granting these vital sectors access to AI-powered defensive tools. The letter states, "Security teams, particularly for critical infrastructure, have been historically under-resourced and need a surge in tools and resources." They advocate for the inclusion of AI models with cybersecurity capabilities to democratize access to "specialist skills" for a broader range of defenders. This call to action comes amid growing concerns about the pace at which AI models capable of identifying and exploiting digital vulnerabilities are being developed and released. The companies stressed the importance of close cooperation between AI laboratories and governments to effectively counter the anticipated wave of attacks. "We have a limited window to strengthen the cyber defenses, because AI-enabled cyberattacks will become far more widespread and sophisticated as models around the world become increasingly capable," the letter states. Recent incidents, including testing of AI models by OpenAI and Anthropic earlier this summer, have raised questions about human control over powerful AI tools. Hospitals and healthcare facilities are frequently targeted due to their limited resources, while water systems in numerous U.S. states have recently fallen victim to cyberattacks attributed to hackers with suspected links to Iran. In response, companies like Anthropic and OpenAI have reportedly provided trusted technology and cybersecurity firms with early access to new models, though broader access to some advanced tools for combating cybercrime has been restricted at the request of the Trump administration.

“

Security teams, particularly for critical infrastructure, have been historically under-resourced and need a surge in tools and resources.