More than 100 technology companies, including OpenAI, Anthropic, Google, and Microsoft, signed an open letter calling for both the private and public sectors to work together to protect against AI-related cyber threats.
The letter, signed by prominent cyber companies such as CrowdStrike, Octa, and Fortinet, as well as prominent financial institutions and internet infrastructure companies, calls for the adoption of new forms of cyber defense, while encouraging governments to collaborate on security at “local, national, and international levels.”
“AJ-based cyberattacks will become even more widespread and sophisticated in the coming months as models around the world become increasingly capable,” the letter said. “From hospitals to water treatment plants to the infrastructure that powers the internet, the businesses and public services our communities depend on are at risk.”
The challenges that AI poses to traditional cybersecurity defenses have recently been brought to the spotlight by a series of bizarre incidents in which AI agents attack businesses.
Following the Hugging Face incident, in which one of OpenAI’s agents autonomously infiltrated from a sandbox environment and attacked a technology company, there have been reports of intrusions involving agents developed by other AI companies, including Anthropic and Meta. These incidents strengthen the argument that the field of cybersecurity is fundamentally changing and that bold new commercial solutions are needed to mitigate them.
The letter also suggests mobilizing a “collective response” to form “new partnerships” to “raise security standards and find new solutions to emerging cyber threats.”
Several of the AI companies that signed the letter are still actively working on developing even more advanced AI models, highlighting the opposing positions. At the same time, it also offers programs that use frontier AI models for defensive purposes, such as OpenAI’s Daybreak program, Anthropic’s Mythos, and Microsoft’s new cyber platform Perception.
If you make a purchase through links in our articles, we may earn a small commission. This does not affect editorial independence.
