The Trump administration has developed an AI cybersecurity framework and shared its details with major artificial intelligence companies including OpenAI and Anthropic on Tuesday, according to WIRED. However, the specifics of this plan remain undisclosed to the public.
According to The New York Times, the framework establishes a voluntary review process focused on evaluating security risks in closed-source AI models. The process will notably exclude AI systems that publish their underlying code, meaning open-source models will not fall under this voluntary review structure.
The framework’s confidential nature has created a situation where major AI laboratories have access to the administration’s security guidelines while the broader public, policymakers, and researchers remain uninformed about the specific requirements and standards being applied to AI development and deployment.