Open-Source AI Leadership Act

This legislation, titled the "Open-Source AI Leadership Act," directs the Secretary of Commerce to actively support the adoption and use of qualified open models in both interstate and foreign commerce. These models are defined as open artificial intelligence models developed by U.S. persons. The Secretary's actions include designating a point of contact, reviewing relevant programs, identifying and addressing barriers to adoption, and entering into agreements with various entities, including private sector companies and foreign partners, to facilitate their use. Furthermore, the bill mandates the Secretary to identify, assess, and publicly disclose risks related to foreign adversary models , which are open AI models developed by covered nations or entities under their control. These risks encompass concerns such as unauthorized training, data confidentiality, organizational security, output accuracy, and potential national security threats like the development of chemical, biological, radiological, or nuclear capabilities. The Secretary must publish annual reports for ten years detailing these risks, providing an overview of leading foreign adversary models, and comparing them with qualified open models. The legislation also clarifies that it does not authorize the Secretary to ban or restrict the availability of open artificial intelligence models. It aims to foster a secure and competitive environment for open AI by promoting U.S.-developed solutions while vigilantly monitoring and informing the public about the potential dangers posed by models originating from foreign adversaries.