Upcoming legislation from Sen. Ted Cruz, R-Texas, looks to establish a federal artificial intelligence sandbox program to evaluate the safety and efficacy of AI technologies, with the effort being overseen by the White House Office of Science and Technology Policy.
In documents reviewed by Nextgov/FCW, the proposed regulatory sandbox program created within OSTP would allow participating AI companies to receive temporary exemption from external AI regulations.
These temporary waivers would be available for “one or more covered provisions of an applicable agency in order to test, experiment, or temporarily provide to consumers artificial intelligence products or services or artificial intelligence development methods on a limited basis without being subject to the enforcement, licensing, or authorization requirements of such covered provisions.”
Part of Cruz’s proposed program, which was first reported by Bloomberg, would create a review process to determine whether an AI software applying for an exemption waiver presents “a health and safety risk, a risk of economic damage, or a risk of unfair and deceptive trade practices.”
The bill further calls for the specific assessment and evaluation criteria to be published in the Federal Register with an open comment period.
Cruz first announced his intent to propose an AI sandbox bill back in May, with the ultimate goal being to remove barriers to AI adoption and prevent overregulation at the state level.
The first draft of his sandbox bill addresses the failed 10-year moratorium that was originally part of the recent budget reconciliation package and would have prohibited the passage of new state-level AI regulation for the next decade by offering waivers on existing regulations for companies testing their AI in the sandbox program.
Federal agencies and private sector companies have participated in sandbox efforts before, such as NVIDIA and non-profit MITRE working to improve and implement AI tools tailored for government workloads.
