In a move to enhance transparency and safety in the rapidly evolving field of artificial intelligence (AI), the Biden administration has initiated a requirement for AI companies to disclose their safety test results to the US government. This significant development was part of an executive order signed by President Joe Biden three months ago, and its implementation is now gaining momentum.
The White House AI Council is set to convene on Monday to review the progress made in implementing this executive order, which has set forth a 90-day goal under the Defense Production Act. This goal mandates that AI companies share critical information, including safety test results, with the Commerce Department.
Ben Buchanan, the White House’s special adviser on AI, emphasized the importance of ensuring AI system safety before their release to the public. Companies have committed to specific categories for conducting safety tests, but a common standard has yet to be established. To address this, the National Institute of Standards and Technology will develop a uniform framework for assessing AI safety, as stipulated in the executive order signed by President Biden in October.
AI has emerged as a paramount concern for both the US economy and national security. The advent of new AI tools, such as ChatGPT, capable of generating text, images, and sounds, has prompted increased government involvement. The Biden administration is also actively collaborating with Congress and international partners, including the European Union, to formulate regulations and guidelines for the responsible management of AI technology.
The Commerce Department has taken steps to draft a rule targeting US cloud companies that provide servers to foreign AI developers, further emphasizing the government’s commitment to overseeing AI development and deployment. Additionally, nine federal agencies, including the Departments of Defense, Transportation, Treasury, and Health and Human Services, have completed risk assessments related to AI’s use in critical national infrastructure, such as the electric grid.
To bolster their regulatory capabilities, the government is actively recruiting AI experts and data scientists across federal agencies. Buchanan emphasized that the aim is not to disrupt the AI landscape but to ensure that regulators are well-prepared to manage this transformative technology.
As the AI industry continues to advance at a rapid pace, this new requirement for AI companies to report safety tests to the US government represents a pivotal step towards ensuring that AI systems are developed and deployed safely and responsibly, benefitting both the public and national security interests.