Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
President Donald Trump He signed an executive order A “voluntary framework” was created for AI companies to share their frontier models with the federal government before they are released to “encourage safe innovation and enhance cybersecurity of critical infrastructure.”
The executive order says the U.S. AI industry has succeeded in part “because we refuse to stifle this innovation through overly burdensome regulation,” but it also recognizes that new AI capabilities come with security risks. Accordingly, it has directed several federal agencies to come up with a framework to “evaluate advanced cyber capabilities of AI models” before rolling them out to the public. Companies will have discretion to share their models with the government’s pre-release, but could obtain some confidentiality protections if they choose to do so. It also requires the federal government to prepare cyber defenses for artificial intelligence, especially for critical infrastructure.
It comes after Trump Postponed at the last minute Signing a pre-planned executive order He was worried It could “stand in the way” of competing with China. While the previous version would have allowed AI companies to voluntarily share their models 14 to 90 days before release, according to New York TimesThe current version requires companies to share their models up to 30 days before public release. Google, Microsoft, and XAI He agreed last month To allow pre-release review by the Department of Commerce’s Center for Artificial Intelligence and Innovation Standards (CAISI).
OpenAI and Anthropy It already agreed to share its models with CAISI back in 2024 under President Joe Biden as part of Biden’s push for Artificial intelligence safety barriers. But until recently, the Trump administration downplayed safety concerns and took a hands-off approach Under the leadership of former White House AI czar David Sachs. The order signed Tuesday explicitly states that it should not be considered a form of mandatory licensing or prior clearance. However, it does reflect some willingness on the part of the Trump administration to employ oversight of AI companies.
“The White House is officially full of legends.”
One factor in the shift may have been Anthropic’s limited rollout in April of its robust Mythos model, which the company said identified “thousands of high-risk vulnerabilities, including some in every major operating system and web browser.” Apparently the legends are as well Create an opportunity to dissolve tensions Between Anthropy and Management, then Legal battle with the Pentagon On its use of artificial intelligence in lethal autonomous weapons and mass surveillance.
So far, the newly signed order has drawn praise even from groups that have defended restrictions on state AI laws. “The White House is officially full of myths,” Brad Carson, president of Americans for Responsible Innovation, said in a statement, adding that the executive order shows the Trump administration takes vulnerabilities in artificial intelligence seriously. Brendan Steinhauser, CEO of the Alliance to Secure AI, said his group is “pleased to see the Trump administration taking the risks of these models seriously,” and both Steinhauser and Carson urged Congress to codify mandatory protections.
