Google, Microsoft and OpenAI Among 100 Firms Demanding Better Cyber Defences

A hundred companies, including Google, Microsoft, Anthropic and OpenAI, have signed an open letter urging governments and organisations to strengthen their cyber defences before AI becomes powerful enough to override them.

"We have a limited window to improve cyber defences," the letter begins. It warns that attacks using AI will become both more widespread and more sophisticated within months, argues that "status quo" security measures "won't be enough," and criticises what it calls the "historic under-resourcing" of security around critical infrastructure.

Who signed

The signatories reach well beyond the AI labs. Banks including Capital One, payment processors Mastercard and Visa, and technology firms Adobe, Oracle and IBM have all added their names — a spread that reflects who actually operates the systems an attacker would target.

The specific ask is that governments supply "capable, defensive AI" and testing to hospitals and water utilities: the operators least likely to have security budgets that match their exposure.

The awkwardness at the centre

The letter's obvious tension is that many of the most capable AI systems were built by the companies now warning about them, and are not all freely available. Anthropic has said its Mythos model can find software vulnerabilities at a level that changes the offensive calculus.

That gives the appeal a double reading. It is a genuine warning from the organisations with the clearest view of what is coming, and it is a request that the public sector fund defences against a capability the private sector is building and selling. Both can be true, and the letter does not attempt to resolve it.