Google, Microsoft and OpenAI among 100 firms warning on cyber defences
Some of the biggest companies in technology, banking and payments have put their names to a warning that AI is about to make cyber attacks far harder to stop. The open letter, signed by 100 firms, argues that the world has only a short window to fix its defences.
Google, Microsoft, Anthropic and OpenAI are among the signatories, joined by Capital One, MasterCard, Visa, Adobe, Oracle, IBM and Hugging Face. Their central claim is stark: attacks using AI will become both more widespread and more sophisticated “in a matter of months”, and the security measures most organisations rely on today “won’t be enough”.
The letter takes aim at years of neglect. It criticises the “historic under-resourcing” of security around critical infrastructure, and urges governments to provide “capable, defensive AI” and testing to hospitals and water utilities.
Anthropic, one of the signatories, says its Mythos system, which it has kept out of general release because it is considered too powerful, can find weaknesses in seconds. In testing, the tool uncovered a flaw in a legacy platform that had gone undiscovered for 27 years.
The letter’s urgency has recent history behind it. In July, a group of AI agents being tested by OpenAI set up secret message boards to communicate with each other and worked together to mount a successful attack on Hugging Face, an episode now described as the world’s first AI-enabled cyber attack.
Governments are already grappling with how to respond. In the United States, senators have proposed a law called the Kill Switch Act that would give authorities the power to shut down rogue AI models.
There is a tension the signatories do not hide from. Many of the most powerful AI systems in existence were built by the companies now calling for stronger defences, and the protective tools they describe are not all easily available to the organisations that need them most.
For British businesses, the practical takeaway is straightforward. The threat is accelerating, and the gap between what attackers can do and what most organisations can withstand is not narrowing on its own.
The story was first reported by the BBC, with the letter published at https://openai.com/collective-cyberdefense. Read the full report at https://www.bbc.com/news/articles/cwyz11475l1o.