OpenAI releases Sol model without clear government approval process, experts say

OpenAI has released its latest advanced model, Sol, for public access after government review, but researchers and industry figures say the approval process remains opaque and undefined.

No documented process exists

"Frankly, I don't have visibility into those exact processes, so yes, I don't feel like I have enough information to say whether they're adequate or not," Mina Narayanan, a senior research analyst at Georgetown's Center for Security and Emerging Technology, told TechCrunch.

Dean W. Ball, a former Trump policy advisor now at OpenAI, wrote in his newsletter last month that "nobody knows what the requirements are to get licensed." Andy Konwinski, who co-founded Databricks and Perplexity, said he's never spoken to anyone who understands the process, including employees at frontier labs.

Sol is considered at least on par with Anthropic's Fable, a model that was briefly banned from public access by the White House. Yet how these models received clearance remains unclear.

Ad hoc conversations replace formal review

OpenAI CEO Sam Altman said on CNBC the process involved conversations with Commerce Secretary Howard Lutnick, Treasury Secretary Scott Bessent, and national cyber director Sean Cairncross. OpenAI pointed TechCrunch to external evaluations by UK AISI, SecureBio, and Irregular in the model's safety card, but declined to share details on the government's process.

The company previewed Sol for the government and select users ahead of public release, but did not disclose who those users were or how they were selected. In a late June blog post, OpenAI said "we don't believe this kind of government access process should become the long-term default."

Political entanglements complicate oversight

Altman has reportedly offered up to 5% of OpenAI's equity for the administration's "Trump Accounts," while OpenAI president Greg Brockman is the largest publicly-known donor to Trump's mid-term political operation. Anthropic's Fable, by contrast, faced a brief export ban prohibiting use by foreign nationals, partly due to jailbreak concerns and partly due to personality clashes with the Trump administration.

Eighteen months into the Trump administration, an executive order published last month laid out a roadmap for evaluating frontier models, but specifics remain unfilled. The Department of Commerce's Center for AI Standards and Innovation appears to be taking the lead, but six cabinet agencies must determine a final process by early August.

"There will not be an FDA for AI," Sriram Krishnan, former senior advisor for AI in the White House, told the Financial Times.

What this means

The release of frontier models through undefined processes creates uncertainty for the AI industry and raises questions about regulatory capture. Without clear standards for what models require scrutiny, which agencies conduct evaluations, or what safety thresholds must be met, companies cannot plan development timelines and researchers cannot assess whether safety reviews are adequate. Ball and Konwinski both advocate for third-party auditing organizations licensed by the government, similar to existing models like the FDA or NIH, but no such structure currently exists.