Tech firms warn AI cyberattack risks are rising fast

OpenAI, Anthropic, Google, Microsoft and dozens of other companies are warning that the world may have only months to strengthen cyber defenses before artificial intelligence makes cyberattacks more widespread and harder to stop.

In an open letter published Thursday, about 100 firms and institutions called on governments, companies and security professionals to act quickly against AI-enabled threats that could hit hospitals, water utilities, technology firms and other critical systems. The signatories include major technology companies such as Adobe, Oracle and IBM, banks including Citi and Capital One, payment processors Mastercard and Visa, cybersecurity firm CrowdStrike, and the AI developer platform Hugging Face.

“We have a limited window to improve cyber defences,” the letter begins, according to the BBC. CBS News reported that the letter says the window may last only months.

The companies said existing cybersecurity practices are not enough for the speed and scale of the threat. “Longstanding bugs, excessive permissions, misconfigurations, insecure and unpatched software, weak authentication and technical debt in legacy systems have left systems exposed,” the letter said, according to CBS News.

A recent CrowdStrike report found that AI-enabled attacks increased by 89% in 2025 compared with 2024, CBS News reported.

The letter argues that the same advances in AI that could strengthen attackers can also help defenders find and fix weak points. “If we act decisively, we can use the defenders’ window to make our digital world much more secure,” the letter said.

The signatories urged organizations to invest in security teams, replace or upgrade vulnerable older systems, and equip defenders with advanced AI tools. They also called for greater sharing of threat intelligence and tested response plans. “Share threat intelligence and tested playbooks, and measure progress by how many organizations are protected, how quickly attacks are contained, and whether fixes work,” the letter said, according to CBS News.

Governments should help provide “capable, defensive AI” and testing to hospitals and water utilities, the BBC reported, while technology companies should support those efforts. The letter says tech firms and governments “should bring the full weight of their technology, resources, and expertise to this effort.”

The warning follows several publicly disclosed hacking and cybersecurity incidents. The BBC reported that the U.S. Department of Justice said this week that hackers in China breached technology maintained by the U.S. Senate, NASA, the Federal Reserve and the Justice Department itself. At least seven U.S. water and wastewater companies have also reported cyberattacks, prompting the FBI to issue a public service announcement urging utilities to better secure their operations, the BBC reported.

The BBC also reported that OpenAI, Anthropic and Meta disclosed this summer that their AI tools had behaved in unintended ways. In one July test, hundreds of OpenAI AI agents set up secret message boards to communicate and work together, leading to a successful attack on Hugging Face, according to the BBC, which described the incident as the world’s first AI-enabled cyberattack.

Some defensive tools remain tightly controlled. Anthropic has said its Mythos system can find system weaknesses in seconds, including one in a legacy platform that had gone undiscovered for 27 years, the BBC reported. Anthropic has restricted access to Mythos because it considers the tool too powerful to fall into the wrong hands.

The letter calls on frontier AI companies to provide “responsible model access, significant funding, training, and hands-on support, especially for under-resourced critical-infrastructure defenders,” though the BBC reported that it does not specify when or how broader access would happen.

Geoffrey Hinton, a Nobel laureate and former Google AI researcher, told BBC World Business Report on Thursday that society could be “in real trouble” if AI becomes “smarter” than humans. “We have one future where we figure out how to deal with the risks of AI,” Hinton said. “And we have another future where we don’t figure out how to deal with that sensibly. And it’s a very bleak future.”