Bipartisan Senators Endorse $32M Annually for AI Research

A bipartisan group of U.S. senators on Wednesday unveiled a road map for artificial intelligence that includes backing a proposal to spend $32 billion annually on civilian research.

Put together by a working group comprising Majority Leader Chuck Schumer, Democratic Sen. Martin Heinrich and Republican Sens. Mike Rounds and Todd Young, the road map aims to detail policy priorities for bipartisan consideration in the Senate.

See Also: Splunk for Public Sector Utilities

The $32 million figures comes from the National Security Commission on Artificial Intelligence, which called on Congress to ensure the federal government spends at least 1% of gross domestic product on scientific AI research. This year, the spending added up to $8 billion.

Most of the non-defense spending will potentially be used in healthcare, Rounds said in a press briefing. Schumer added that an “equal amount” would likely be sought for defense research.

The road map does not take a prescriptive approach to developing AI policy Schumer said, suggesting that Congress take a piecemeal approach to legislation rather than coming up with one comprehensive bill.

“Whatever changes AI brings, it isn’t going to discriminate between the left and right,” Schumer added, addressing the need for bipartisan action.

The road map focuses on eight areas including innovation, workforce use cases, elections and democracy, privacy and liability, transparency, explainability, intellectual property and copyright, and AI risk prevention and national security.

It calls for the development of a framework to determine when an AI system should be subject to additional evaluation, rather than mandating the process for all upcoming tools. It also does not suggest revamping copyright law to address the ongoing battles in the area but instead asks that policymakers consider whether new legislation on transparency, content provenance and likeness protection is necessary.

The 31-page report does not detail specific legislations – just guidelines and considerations on which legislators can base AI laws on.

The road map carries “strong ‘I had to turn in a final report to get a passing grade so I won’t think about the issues and will just copy old stuff and recycle it’ vibes,” said Suresh Venkatasubramanian, co-author of the White House’s AI Bill of Rights. The most “glaring” issue, he said, is the lack of a rigorous approach to governing rights that affect AI systems – “a complete whiff on investments in sociotechnical research needed to make sure we can govern AI systems effectively,” he said.

The senators held nine forums with AI policy leaders in the fall of 2023, including participation from industry executives such as OpenAI’s Sam Altman and Google’s Sundar Pichai, along with academics, labor and civil rights leaders.

The road map also addresses the “black box” nature of how an AI system functions, which may raise questions about whether companies abide by existing laws. New regulatory language must allow regulators to “access information directly relevant to enforcing existing law and, if necessary, place appropriate, case-by-case requirements on high-risk uses of AI, such as requirements around transparency, explainability, and testing and evaluation,” it said.

Forum participants discussed “at length” the use of watermarking as a means to distinguish AI-generated content from human content. Watermarking is a favored buzzword of both tech giants and the White House.

The road map is so far “the most comprehensive and impactful bipartisan recommendations on artificial intelligence ever issued by the legislative branch,” Young said.