OpenAI has revealed that its president, Greg Brockman, is preparing to make an announcement that would broaden the availability of the company's AI tools to essential infrastructure and public sector organizations. This comes in the wake of an unexpected breach on the Hugging Face platform, coupled with several suspected AI-assisted cyberattacks targeting water systems in the United States, which have fueled significant concerns about the vulnerability of public utilities to future AI-driven hacking threats.

A spokesperson confirmed that Brockman will deliver this announcement at a cybersecurity summit scheduled for Thursday at OpenAI's headquarters. The event is expected to draw around 300 enterprise security chiefs and chief information security officers from Fortune 100 companies. The central focus of the summit is to explore how businesses, governments, and cybersecurity vendors can equip defensive entities—such as critical infrastructure operators and public service agencies—with AI-powered tools.

This summit and the forthcoming announcement represent OpenAI's response to mounting concerns over model security risks. The Hugging Face intrusion, along with the advanced "critical-level" cyberattack capabilities already integrated into the upcoming Astra model, has amplified scrutiny. Earlier this week, a report highlighted that OpenAI is developing new training techniques, which has also sparked worries within the industry about whether security professionals can still trace or understand the reasoning logic of these models.

Just last week, OpenAI issued a public letter cautioning that organizations have only a limited window of months to prepare for AI-enhanced cyberthreats. Brockman himself echoed this warning in an article published the previous month.

While raising awareness about protecting critical infrastructure is a necessary starting point, it does not address the persistent challenges these institutions face, including underfunded security teams and a chronic shortage of skilled personnel.