Microsoft, Google, xAI agree to share AI models with White House for security reviews
Microsoft, Google’s DeepMind, and xAI, backed by Elon Musk, have agreed to share their advanced AI models with the US government for security reviews, as announced by the Department of Commerce.
The Center for AI Standards and Innovation will be conducting pre-deployment evaluations and targeted research to assess the capabilities and risks associated with these cutting-edge technologies.
Agreements with Anthropic and OpenAI have also been updated to align with new directives on security reviews from Commerce Secretary Howard Lutnick and President Trump.
Concerns have been raised about the potential misuse of AI tools like Anthropic’s Mythos, which could lead to cyberattacks and other security threats if misused.
Google DeepMind has confirmed the partnership, while other parties involved have not responded to requests for comments.
Chris Fall, the director of the Center for AI Standards and Innovation, emphasized the importance of independent measurement science in understanding the implications of frontier AI on national security.
Industry collaborations are expanding to address the challenges of regulating AI technologies in the public interest.
The establishment of the AI Safety Institute, renamed as CAISI under the Trump administration, aims to enhance AI safety protocols to facilitate the deployment of new AI models.
Despite efforts to accelerate AI development, concerns remain about the potential risks associated with AI technologies like Anthropic’s Mythos.
Anthropic’s Mythos has been found to have vulnerabilities that could be exploited in critical infrastructure, leading to security threats if not properly managed.
OpenAI plans a limited release of its latest model, GPT-5.5-Cyber, due to security concerns.
Public opinion on AI innovation is mixed, with a Pew Research Center poll indicating equal levels of concern and excitement about the increased use of AI in daily life among Republicans and Democrats.
