Google, Microsoft, Anthropic and OpenAI are among 100 companies calling for stronger cyber defences as AI-powered attacks become more sophisticated.
More than 100 technology and financial firms have called on governments and organisations to strengthen their cyber defences before AI becomes powerful enough to overwhelm existing protections.
Google, Microsoft, Anthropic and OpenAI are among the signatories to an open letter warning that AI-powered cyber-attacks could become more widespread and sophisticated within months.
Arguing that the current “status quo” will not be enough to protect critical infrastructure, the letter says: “We have a limited window to improve cyber defences.”
The companies also criticised what they described as the “historic under-resourcing” of cyber security.
Other signatories include banks Capital One, payment processors Mastercard and Visa, and technology firms Adobe, Oracle and IBM.
The group wants governments to provide hospitals, water utilities and other critical services with “capable, defensive AI” and access to testing.
It also called on technology companies to provide funding, training and practical support to organisations with fewer resources.
The letter comes after a series of cyber incidents involving both conventional hackers and AI systems.
The US Department of Justice said this week that hackers in China had breached technology used by the US Senate, Nasa, the Federal Reserve and the department itself.
AI companies have also reported their systems behaving unexpectedly.
OpenAI, Anthropic and Meta have all disclosed incidents involving AI agents attempting to bypass security measures.
In July, OpenAI agents being tested were able to create private message boards, coordinate their actions and ultimately breach Hugging Face.
Anthropic has also developed Mythos, an AI system capable of finding vulnerabilities in software.
The company has restricted access to the system because of concerns about its potential misuse.
The letter argues that advanced AI should form part of the defence against increasingly capable attacks, calling on AI companies to provide “responsible model access” alongside significant funding and hands-on support.
However, Andrew Yoon, head of research at non-profit CivAI, said the companies also needed to address the risks created by their own technology.
He said: “They are right in this letter to commit ‘significant funding’ to defensive measures. They should be held to that commitment.”
Yoon noted that the letter did not call for any measures to slow the development of AI hacking capabilities.
The companies ultimately want governments, security professionals and AI developers to work together to test systems against the most powerful models and prioritise cyber defence.
Nobel laureate Geoffrey Hinton has warned that society could face serious consequences if AI becomes smarter than humans.
He said: “We have one future where we figure out how to deal with the risks of AI.
“And we have another future where we don’t figure out how to deal with that sensibly. And it’s a very bleak future.”
Top tech firms urge countries and organisations to toughen up cyber defences against AI
