OpenAI published a request for mandatory, capability-based national AI safety requirements and said it would continue supporting state legislation while Congress considers a federal framework. The company backed four California bills covering independent assessments, auditor standards, youth safeguards and biological-risk protections. This account draws on Associated Press and Axios and distinguishes confirmed events from statements, allegations, proposals, or preliminary findings.
The available record in Editorial: AI’s Governance Window Is Finally Concrete contains several additional specifics: Jacob Coxon resigned from Anthropic after working in pretraining research at both Anthropic and OpenAI and publicly criticized competitive pressure between frontier laboratories. Coxon’s most severe predictions remain his stated judgment, not an independently established forecast or evidence of a specific imminent event. OpenAI separately described work on a serious-incident reporting policy, illustrating that concrete governance can be specified without resolving every dispute about long-term risk. These details set out the sequence and the measurable actions reported by the cited sources.
Relevant factual context includes the following: Mandatory evaluations require definitions of covered capabilities, independent access, reporting deadlines and enforcement consequences. Protected internal escalation can preserve evidence from employees while avoiding the assumption that every employee claim is correct. Federal and state systems can complement or conflict with one another depending on preemption, agency authority and shared technical standards. That background helps locate the new development in an existing legal, operational, scientific, or civic process without extending the evidence beyond what the sources support.
The record is not complete. No binding national framework had been enacted, and the available public record did not validate Coxon’s timeline or show how OpenAI’s proposed thresholds would operate in statutory text. The next scheduled or observable markers are legislation translating capability-based requirements into testable rules and publication of incident-reporting standards and independent evaluation results from frontier laboratories Any later change in the underlying facts belongs in a subsequent update rather than being assumed here.