OpenAI CEO Sam Altman stated that his company made "many changes" as it engaged in talks with White House officials ahead of debuting its newest AI systems to the public. The discussions occurred over several weeks and involved multiple government agencies.
During a Thursday interview with CNBC, Altman mentioned that the company behind ChatGPT experienced a "collaborative back and forth" with senior U.S. officials, among them Treasury Secretary Scott Bessent and Commerce Secretary Howard Lutnick, throughout the period preceding the broad launch of GPT-5.6.
The original rollout of this model family was restricted to a few chosen collaborators after the Trump administration urged a phased deployment. On Thursday, OpenAI will make GPT-5.6 accessible to everyone, following a go-ahead from U.S. government authorities, the company confirmed. The green light came after OpenAI provided assurances about safety and security.
Anthropic PBC had to temporarily disable two of its AI systems, Fable 5 and Mythos 5, following export restrictions imposed by the Commerce Department aimed at blocking foreign access to those models. Those measures were reversed at the end of last month after several weeks of talks, during which Anthropic added extra cybersecurity protections for its systems. The quick reversal of the export controls suggests that the government is willing to accommodate companies that address its security concerns. The episode with Anthropic demonstrates that even rivals face similar regulatory hurdles.
Get the market news that matters in a five-minute read with Market Briefs, our free daily newsletter
The negotiations highlight the increasing role of the federal government in shaping the deployment of advanced AI systems. The Trump administration has been vocal about ensuring that AI models do not pose national security risks, leading to close scrutiny of releases by major companies like OpenAI and Anthropic. Both companies have had to adapt their strategies to comply with regulatory expectations.
Altman's positive view of the process suggests that these regulatory interactions, while initially challenging, may become standardized, easing future rollouts.
The process involved multiple rounds of testing and feedback, which both sides found productive. The discussions also underscore the delicate balance between fostering innovation and protecting national interests, a challenge that will only grow as AI capabilities expand.
Altman refrained from specifying the modifications OpenAI undertook in the talks, though he noted that the government is now evaluating the new models and checking for any issues.
"I think that's good as long as the process is understandable, fair and quick," he said. "The government seems to really share those goals and my guess is our next time through, our next model, it'll be much smoother. We understand it and know how to go with it better."
"Asked about whether OpenAI intends to move forward with a public offering this year, Altman replied", "I don't know."
Altman's hope for smoother future releases indicates a maturing relationship between AI developers and regulators.
Join Market Briefs, our free daily newsletter, for a quick daily rundown of the markets