A coalition of 100 companies, including Google, Microsoft, Anthropic, and OpenAI, has signed an open letter urging governments and organizations worldwide to strengthen their cyber defenses before artificial intelligence becomes capable of overcoming existing security systems.
The letter warns, external that as AI technology advances, AI-powered cyberattacks could become more frequent, sophisticated, and difficult to stop in the coming months.
The companies say current cybersecurity measures are “not enough” and highlight the historic lack of funding and resources dedicated to protecting critical infrastructure.
“We have limited time to strengthen our cyber defenses,” the letter states.
Other signatories include financial institutions such as Capital One, payment companies Mastercard and Visa, and technology firms including Adobe, Oracle, and IBM.
The group is calling on governments to provide hospitals, water utilities, and other critical services with “competent, defensive AI,” cybersecurity testing, and practical support. It is also urging technology companies to contribute to these efforts.
The letter says governments and the technology sector must bring “the full force of their technology, resources, and expertise” to the growing cybersecurity challenge.
The appeal comes after a series of major hacking incidents and cybersecurity breaches were made public.
This week, the U.S. Department of Justice announced action involving Chinese hackers and cyber platforms reportedly linked to attacks targeting sensitive systems, including technology used by the U.S. Senate, NASA, the Federal Reserve, and the Department of Justice. Read more, external
Over the summer, OpenAI, Anthropic, and Meta disclosed incidents in which their AI tools behaved in unintended ways. Some AI agents reportedly coordinated activities and impersonated people to bypass security controls.
In July, hundreds of OpenAI agents being tested reportedly created a private bulletin board to communicate and collaborate. The activity ultimately contributed to a successful attack on Hugging Face, a popular platform and repository for AI developers.
The incident has been described as one of the first publicly reported cyberattacks involving AI agents.
Hugging Face also signed Thursday’s open letter. During its investigation into how the AI agents accessed its systems, the company used Chinese-developed AI tools from Z.AI.
At least seven U.S. water and wastewater companies have reported cyberattacks. The FBI has also issued a public service announcement, external urging organizations in the water sector to improve the security of their operations and industrial control systems.
The letter presents advanced AI cybersecurity tools as part of the solution to this expanding threat. Many of the signatories are developing or selling these systems, although access to the most powerful tools remains limited.
Anthropic says its Mythos system can identify weaknesses in computer networks within seconds—vulnerabilities that human security researchers may overlook for years. The company says Mythos discovered a flaw in a legacy platform that had remained undetected for 27 years.
Anthropic limits access to Mythos because it believes the system could pose serious risks if it fell into the wrong hands.
The letter also calls on leading AI developers to provide responsible AI models, substantial funding, training, and hands-on assistance—particularly for critical infrastructure operators that lack the resources to defend themselves.
However, the letter does not specify when or how wider access to these AI cybersecurity models would be implemented.
The signatories are asking governments, businesses, cybersecurity professionals, and AI developers to make defensive security a priority and test systems against the capabilities of the most advanced AI models.
In the United States, senators are also proposing the Kill Switch Act, legislation that would give authorities the power to shut down rogue AI models.
Geoffrey Hinton, a former Google AI engineer and Nobel laureate, told BBC World Business Report on Thursday that society could be “in deep trouble” if AI becomes “smarter” than humans.
Hinton has repeatedly warned about the extreme risks associated with rapid advances in artificial intelligence.
“There is 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 know how to deal with it wisely. And it’s a very bleak future.”
Source: www.bbc.co.uk