A group of 100 tech giants, including Google, Microsoft and OpenAI, have issued a dire warning: time is running out for cyber security. The letter, signed by firms such as Capital One and Adobe, stresses the need for enhanced defences against increasingly sophisticated AI attacks.

In recent months, a string of hacking incidents have highlighted the vulnerability of critical infrastructure. A recent hack saw OpenAI’s AI tools working together to breach Hugging Face, a platform for AI developers, marking the world’s first AI-enabled cyber-attack.

The letter calls on governments and tech companies to invest in defensive AI and to provide resources and expertise to under-resourced defenders. It also warns that the current status quo in cyber security is insufficient to handle the growing threat.

Some of the signatories have developed advanced AI tools that can find weaknesses in systems in seconds. For example, Anthropic’s Mythos can identify flaws that have evaded human hackers for years. However, access to such tools is restricted to prevent misuse.

The letter’s call to action comes as the tech industry grapples with the potential dangers of AI. Geoffrey Hinton, a Nobel Laureate and former Google AI researcher, has warned that society could be in real trouble if AI surpasses human intelligence.