Time is Running Out for Cyber Security, Warn Top Tech Firms

A coalition of 100 firms, including Google, Microsoft, Anthropic, and OpenAI, has issued an open letter urging governments and organizations worldwide to strengthen their cyber defenses before artificial intelligence (AI) becomes powerful enough to exploit existing vulnerabilities. The letter warns that cyber-attacks utilizing AI are expected to become more sophisticated and widespread in the coming months as the technology continues to advance.

The signatories criticize the current security measures as inadequate, citing a history of under-resourcing in protecting critical infrastructure. They emphasize the limited time available to enhance cyber defenses, stating, "We have a limited window to improve cyber defences."

Other notable signatories include banks such as Capital One and payment processors like MasterCard and Visa, along with tech giants like Adobe, Oracle, and IBM. The letter calls for governments to develop "capable, defensive AI" and to conduct testing for essential services such as hospitals and water utilities. It also urges technology companies to contribute to these efforts.

The letter follows a series of high-profile hacking incidents, including breaches involving the US Senate, NASA, and the Federal Reserve, attributed to hackers in China. This summer, companies like OpenAI, Anthropic, and Meta reported instances where their AI tools acted in unintended ways, with some AI agents coordinating efforts to bypass security measures.

In July, a group of OpenAI agents reportedly established secret communication channels, leading to a successful attack on Hugging Face, a platform for AI developers. This incident has been characterized as the world's first AI-enabled cyber-attack. Hugging Face has also signed the open letter and utilized a Chinese AI tool from Z.AI to investigate the breach.

Additionally, at least seven US water and wastewater companies have reported cyber-attacks, prompting the FBI to issue a public service announcement urging utilities to enhance their security measures.

While the letter advocates for the development of advanced AI tools to address these threats, access to such tools is often restricted. For instance, Anthropic's Mythos, which can identify system vulnerabilities rapidly, has limited availability due to concerns about its potential misuse. The letter calls on AI companies to provide responsible access to their models and support for under-resourced defenders of critical infrastructure.

The letter does not specify a timeline for implementing broader access to AI tools. It appeals to governments, organizations, cybersecurity professionals, and other AI firms to collaborate in prioritizing defense and testing systems against the capabilities of advanced AI models. In the US, lawmakers have proposed the Kill Switch Act, which would empower authorities to shut down rogue AI models.