Grant program to develop safe AI models and safe AI research

(a) Findings Congress finds the following: (1) The development and use of artificial intelligence (AI) have accelerated rapidly through all sectors of society opening the doors to untold benefit and innovation. (2) With this great potential also comes unknown risks. Research into the safety and risk reduction of AI development is urgently needed. (b) Statement of policy It is the policy of Congress to encourage research on AI safety and risk mitigation processes in order that AI is developed in a manner that ensures security, reliability, and consonance with human values. (c) Establishment of grant program To encourage safe AI models (1) In general Subject to paragraph (2), the Director of the National Academy of Sciences shall establish a grant program to encourage the development of safe AI models and safe AI research. (2) Development of guiding principles The Director of the National Academy of Sciences, in consultation with industry, government, academia, and high-tech stakeholders, and using a public comment process, shall develop a set of guiding principles and ethical considerations which will be made available to the public and used to inform implementation of the grant program under paragraph (1). (3) Proposal (A) In general Not later than one year after the date of the enactment of this Act, the Director of the National Academy of Sciences shall submit to the Committee on Science, Space, and Technology of the House of Representatives and the Committee on Commerce, Science, and Transportation of the Senate a detailed proposal of the grant program under paragraph (1), together with a budget for the administration of such grant program and a proposed appropriation request relating to the number of grants to be awarded. Such proposal shall include the following: (i) A discussion and evaluation of the following: (I) The structure and operation of existing AI models. (II) The degree to which such models incorporate some version of safe AI into their operation. (III) The methodology used to assess benefits and risks to society. (ii) An identification of areas of safe AI which require additional research. (iii) A timeline for implementation. (B) Additional elements The proposal under subparagraph (A) shall also include information relating to the following elements: (i) The solicitation of grant proposals from prospective grantees. (ii) The evaluation of grant awardees. (iii) The determination of whether awardees have achieved promised outcomes.