OpenAI has confirmed it will comply with President Trump’s executive order requiring AI companies to submit their models for government review before public release. The decision signals the company’s willingness to work with federal authorities on AI safety oversight as the technology becomes increasingly powerful.
George Osborne, OpenAI’s head of countries, told CNBC that the startup would participate in the voluntary program. “It’s quite right that democratic governments have a big role to play in how this technology is used and deployed,” he said during an interview at SXSW in London.
The executive order, signed by Trump on Tuesday, requires AI companies to provide access to their models 30 days before release. The government will then conduct benchmarking tests to assess the models’ capabilities and determine which ones qualify as “covered frontier models” based on their advanced cyber capabilities.
This move reflects growing government concern about AI safety as models become more sophisticated. The Biden administration had previously pushed for similar oversight measures, but Trump’s order takes a more structured approach by establishing specific timelines and assessment criteria.
Osborne, who served as the UK’s finance minister from 2010 to 2016, emphasized that OpenAI takes its safety responsibilities seriously. “As this leading frontier lab with these very, very powerful and capable AI models, we don’t wait to be asked,” he said. “We proactively suggested ways that governments can keep track on safety and security issues, not just in the U.S., but more broadly.”
The compliance decision comes at a critical time for AI regulation. Tech companies are racing to develop increasingly powerful models while governments worldwide struggle to establish appropriate oversight frameworks. OpenAI’s cooperation could set a precedent for how other major AI companies respond to regulatory requirements.
For the AI industry, this represents a shift toward more formal government oversight. While the current order is voluntary, it establishes infrastructure that could easily become mandatory if safety concerns escalate. Companies that participate early may have more influence over how future regulations are shaped.
Osborne noted that governments need to be strategic in their regulatory approach. “What we suggest to governments is they create powerful regulatory bodies, but with a lot of flexibility into how they will operate in the future,” he said. This reflects the tech industry’s preference for adaptive regulation that can evolve alongside rapidly advancing technology.
The 30-day review period will likely test both government capacity to assess complex AI systems and companies’ ability to predict potential risks before release. The process could also slow down AI development cycles, potentially affecting the competitive landscape as companies factor review time into their release schedules.
