U.S. President Donald Trump disembarks from Air Power One upon arrival from Miami on Could 3, 2026 at Joint Base Andrews, Florida.
Roberto Schmidt | Getty Pictures
The Middle for AI Requirements and Innovation on Tuesday announced agreements with Google DeepMind, Microsoft and Elon Musk’s xAI that can enable the U.S. authorities to judge artificial intelligence fashions earlier than they’re publicly accessible.
CAISI, which sits underneath the U.S. Division of Commerce, will “conduct pre-deployment evaluations and focused analysis to higher assess frontier AI capabilities and advance the state of AI safety,” in accordance with a launch.
The announcement builds on CAISI’s previous partnerships with OpenAI and Anthropic from 2024, the middle stated. These agreements have been renegotiated to mirror CAISI’s directives from the Commerce Secretary Howard Lutnick and America’s AI Motion Plan, in accordance with the discharge.
Past CAISI’s announcement on Tuesday, the White House has been weighing the creation of a brand new AI working group that may discover potential oversight procedures, together with plans to vet fashions earlier than they’re launched to the general public, CNBC confirmed.
The working group would carry collectively quite a lot of tech executives and authorities officers, in accordance with a supply near the discussions who requested to not be named as a result of the small print are confidential. The group could also be established by an govt order.
The White Home instructed CNBC that dialogue about potential govt orders is concept, and that any coverage announcement will come immediately from President Donald Trump.
The New York Times was first to report the working group.
The White Home’s discussions a few working group come after Anthropic, a number one AI firm, caught authorities officers’ consideration final month by asserting a robust new mannequin, Claude Mythos Preview.
Mythos excels at figuring out weaknesses and safety flaws inside software program, in accordance with Anthropic, so the corporate determined to restrict the rollout to a choose group of corporations as a part of a brand new cybersecurity initiative known as Venture Glasswing.
Anthropic CEO Dario Amodei met with senior members of the Trump administration to discuss Mythos at the White House simply days after it was introduced. That assembly occurred even after the Protection Division had declared Anthropic a supply chain risk, that means the corporate purportedly threatens U.S. nationwide safety.
Each the White Home and Anthropic characterised that assembly as “productive.”
WATCH: White House reportedly considering vetting AI models before they are released
