WASHINGTON — Microsoft President Brad Smith called for clearer artificial intelligence (AI) regulation during an interview on the sidelines of the AI for Good Global Summit on July 9, 2026. Smith criticized the U.S. government's use of export controls as a primary method to restrict the release of AI models.

"Everyone is reluctant to say there should be regulation, but what we really have right now is regulation without transparent or complete rules," Smith said. He stated that the Commerce Department invoked export-control law to force Anthropic to pull its Fable 5 and Mythos 5 models from the market worldwide in June 2026, citing a cybersecurity risk.

"They asked Anthropic to take Fable off the market," Smith said. "Anthropic said it would not, so they used an export control lever in a way that caused Anthropic to take it off the market, both domestically and internationally." Anthropic's Fable 5 model became available again in early July 2026.

Weeks after the restrictions concerning Anthropic, officials pressed OpenAI to hold back the public rollout of its GPT-5.6 model family. OpenAI limited early access to the GPT-5.6 model to government-vetted partners. OpenAI stated that the GPT-5.6 model would launch publicly on Thursday, July 9, 2026.

Smith acknowledged the government's reasons for concern regarding national security. "The U.S. government got information that led it to conclude that there was an urgent cybersecurity risk, and when the government gets that information, I think it's right to act," he said. However, Smith indicated that relying solely on export controls demonstrates a lack of appropriate regulatory tools. "But, what the government found was that it only had one regulatory tool it could use: an export control tool," Smith said. "The government doesn't have the tools it needs."

Smith emphasized that businesses cannot plan without clear guidelines. "Without rules, businesses can't plan," he said. He added that companies require certainty regarding market and supply access. "We want to sell our services around the world, but people will not buy what we have to sell unless they're confident that there will be certainty of supply, continuity of supply, and we need to address that," Smith said.

In June 2026, an executive order established a voluntary pre-release review process for AI models but explicitly avoided creating a formal licensing system for frontier AI models. The government has not disclosed the criteria for who qualifies as a "trusted partner" for AI model access, nor has it disclosed which AI models will be subject to vetting in the future.