Meta $META is the only major U.S. artificial intelligence developer that has yet to reach an agreement with the Trump administration to voluntarily submit new AI models to federal officials for security reviews before they are released, according to The New York Times.
Citing unnamed sources, The Times reports that officials are pressuring Meta to join OpenAI, Anthropic, Google $GOOGL, xAI, and Microsoft $MSFT in reaching a deal to provide the Center for AI Standards and Innovation with advance access for national-security evaluation purposes.
Meta spokesperson Francis Brennan told Te Times that the company shares the administration's goal of advancing U.S. leadership on "robust and secure frontier AI" and expects to reach an agreement. "While we are working through the details, we hope to sign the agreement soon," Brennan said.
Commerce Department spokesman Ben Kass confirmed that the Center for AI Standards and Innovation, which is housed in the department and overseen by Commerce Secretary Howard Lutnick, engages companies about voluntary agreements.
An executive order President Donald Trump signed on June 2 created a voluntary structure giving federal officials a window of as many as 30 days to assess AI models ahead of any release to trusted partners. The order also called for development of a classified benchmarking process to assess advanced AI capabilities. It does not create any mandatory licensing or preclearance requirement for AI models.
Meta released its most recent AI model, Muse Spark, in April. In contrast to the open-source Llama line Meta had long been known for, Muse Spark keeps its underlying code proprietary.
The administration's push for model reviews comes as Washington's concerns about AI security have grown.