OpenAI Missives Spur Debate Over Who Shoulders AI Cyber Defenses

OpenAI Inc. wants companies to spend more on cyber defense and use its services to fend off rising threats posed by sophisticated AI models — like the ones the tech giant itself released.

For many businesses, that’s a tough ask.

OpenAI’s call for collective action against AI-enabled cyber attacks absolves the firm of some of its responsibility and puts the onus on companies already struggling with tight budgets and outdated technology, cyber professionals said. And if companies neither act themselves nor get more support from OpenAI and other AI developers, they face rising liability exposure. The rising risk was highlighted by Anthropic and other global AI developers themselves on Saturday as they called for a slowdown on tech development.

“Effectively, this is a problem they more or less created,” said Greg Notch, chief technology officer at security services provider Expel and former head of information security at the NHL. And “injecting fear, uncertainty and doubt at the board level” is a “tried and true” method to unlock budgets, he added.

More than 100 tech companies — including Anthropic PBC, Google, Microsoft Corp., and CrowdStrike Holdings Inc. — signed an open letter late last month from OpenAI suggesting businesses and their vendors should unite to fix network weaknesses. About a week later, OpenAI announced a $1 billion pledge to help utilities, local government, and financial institutions strengthen their defenses, all while reiterating its message that advanced AI can “help defenders move faster.”