Get Full Government Meeting Transcripts, Videos, & Alerts Forever!
Get email alerts on the Defensive Support topic
No spam. Unsubscribe anytime.
OpenAI tells lawmakers it can support California defenders with credits and specialized access
Summary
OpenAI said it welcomes California's new AI cyber defense program, offered credits and vetted access for defensive evaluations, and described Daybreak, its program to embed cyber-specific models into security workflows for vetted defenders.
OpenAI told the committee that it stands ready to support California's planned AI cyber defense program and has offered a mix of access, training and credits to speed defensive adoption.
John Lindsay described Daybreak, OpenAI's program to embed frontier AI capabilities into secure software-development and incident-response workflows. He said the company has expanded Daybreak and introduced cyber-specific models to support vulnerability research and security testing for approved defenders. "We sent officials in several states, including California, a letter offering up to $1,000,000 in no cost API credits for controlled authorized cyber evaluations," Lindsay said, adding that OpenAI "stands ready to support the implementation of this program."
Panelists and state officials said the credits, combined with verification programs and training, are meant to get advanced defensive capabilities into the hands of vetted security teams and state partners while limiting access for unvetted actors. Jonathan Snow (Cal OES) described CalSecure 2.0 and the state's secure portal for SB 53 reporting; he said the state is building a playbook for collaborative incident responses and coordinating with federal and private partners.
