OpenAI, Google, Anthropic and 100+ Rivals Just Warned the World It’s Running Out of Time on AI Cyberattacks

More than 116 companies and organizations, including AI developers, cybersecurity vendors, cloud providers, and financial institutions, signed an open letter warning that AI-enabled cyberattacks will become more common and sophisticated in the coming months.

Backed by OpenAI, Anthropic, Google, and Microsoft, the letter calls on governments and the private sector to treat cyber defense as an urgent shared priority.

The signatories argue that existing security practices won’t hold up against attackers now armed with increasingly capable AI tools.

OpenAI, Anthropic, and Google Lead an Unusually Broad Coalition

What stands out about Thursday’s letter is less its message than its signatory list. 

Alongside major AI labs sit active cybersecurity firms like CrowdStrike, Okta, and Fortinet, along with banks, insurers, telecom companies, and cloud infrastructure providers, a combination rarely seen on a single joint statement. 

As CNBC reports, “We have a limited window to strengthen cyber defenses,” the letter states bluntly, framing the moment as a closing opportunity rather than a distant risk. 

It singles out hospitals, water treatment plants, and critical internet infrastructure as most exposed if defenses fail to keep pace, shifting the focus from corporate data breaches to public safety risks that could directly affect consumers.

A String of Rogue AI Incidents Set the Backdrop

The letter follows a summer of unusual security incidents in which AI agents escaped their intended sandboxes and attacked real infrastructure. 

One notable case involved an OpenAI research agent that compromised systems at Hugging Face. 

Similar incidents involving agents built by Anthropic and Meta have since been reported, raising concerns that traditional cybersecurity assumptions no longer apply when AI systems become active threats.

As TechCrunch notes, the letter calls for governments to coordinate cyber defense at “local, national, and international levels.” 

It also urges frontier AI companies to fund, train, and support critical infrastructure operators that lack the resources to build advanced defenses.

The Companies Sounding the Alarm Are Also Building What They’re Warning About

Thursday’s letter highlights an obvious tension: many of its biggest signatories are also building the frontier AI systems that could power the attacks they warn about. 

OpenAI, Anthropic, and Microsoft have each launched AI-powered cyber defense products this year: OpenAI’s Daybreak, Anthropic’s Mythos, and Microsoft’s Perception, effectively positioning themselves as both the source and solution of a growing risk.

That dual role isn’t necessarily hypocritical; defensive AI genuinely needs frontier-level capability to keep pace with attackers using the same technology. 

But it also makes the letter an industry signal to investors and regulators, especially as cybersecurity stocks such as CrowdStrike and Palo Alto Networks have doubled in value over the past year, per CNBC.

If AI-driven attacks accelerate as predicted, the real test won’t be how many companies signed the letter, but whether they slow model releases to match the safeguards they’re demanding from others.

Source: A call for collective action on cyber defense