First crack at comprehensive AI legislation coming early 2024 from Senate Commerce Chair Cantwell

Senate Commerce Committee Chair Maria Cantwell is readying a series of significant bipartisan bills related to artificial intelligence, including efforts to balance the regulation of popular generative AI tools as well as initiatives to boost innovation, making it the first true comprehensive legislation in Congress to tackle the issue of AI.

Cantwell, a Democrat from Washington, is expected in the coming weeks to introduce the legislation with a series of bills related to relevant AI issues like deepfakes, jobs and training, algorithmic bias, digital privacy, national security, and AI innovation and competitiveness, according to Cantwell’s staff and four sources familiar with the legislative effort.

The comprehensive series of AI bills has the support and blessing of Senate Majority Leader Chuck Schumer, D-N.Y., who has tapped multiple Senate committee chairs to lead on introducing and debating major AI legislation after the culmination of his bipartisan AI Insight Forums last year, three sources familiar with the legislative effort told FedScoop.  

“The AI bills won’t come out all at the same time; they’ll be dropped in a series, in a staggered fashion, but we’re aiming for the next few weeks and months as soon as possible,” a senior legislative aide for the Senate Commerce Committee majority staff told FedScoop. “It’s a top priority for the senator especially because other countries and the U.S. need to be ahead on AI policy and AI competitiveness.

“Senate Commerce has the primary or at least very important jurisdiction on AI policy and a majority of AI policy is already coming out of our committee. Many bills have been referred to us, so we want to build upon that and work with Republicans to put out something that can move,” the senior aide added.

Cantwell announced at various points in 2023 that she’s working on introducing AI-related bills, including legislation on threats posed by deepfakes, a federal privacy bill targeting AI discrimination, a reskilling “GI bill” for AI, as well as legislation on potential disruptions to jobs and education posed by AI. 

She has yet to actually introduce any AI legislation, but has made it a priority for herself and the Senate Commerce Committee in the next few months. 

Two AI scholars familiar with legislative efforts in Congress told FedScoop that they expect Cantwell’s comprehensive AI legislation to start with the introduction of bills that focus on a few areas of shared bipartisan interest.

“The low-lying fruits are AI bills related to deepfakes in a narrow fashion, AI research and development, consumer fraud, and workers displaced by AI,” said Samuel Hammond, a senior economist focused on AI policy at the Foundation for American Innovation, a tech-focused, libertarian-leaning think tank previously known as the Lincoln Network.

“The vibes are there is some agreement but nothing that’s clearly going to go all the way,” Hammond added. “It wouldn’t surprise me given this is the Commerce Committee that Cantwell uses the bills to follow up on the CHIPS and Science Act, to get the most bang for their buck.”

Daniel Colson, the founder and executive director of the AI Policy Institute, said that he expects Cantwell’s series of comprehensive AI bills to focus first on bias and discrimination caused by AI, followed by legislation to address the most displaced workers, like language translators. There could also be bills to regulate the most extreme risks that large AI models that cost $10 billion or more could bring, he said. 

Three AI scholars familiar with Cantwell’s AI legislative efforts said the legislation could include a spending package related to AI policymaking between $8 billion and $10 billion.

Gathering bipartisan momentum for any major AI legislative effort has proven challenging, given the chasm between Democrats and Republicans in Congress and within the Senate Commerce Committee in particular. 

“Republican Commerce Committee staff said at a meeting with some of us recently that ‘we’re just going to hold the line’ on all AI-related legislation,” a senior AI scholar who met with Senate Republican Commerce staffers at the end of 2023 told FedScoop. The source added that Commerce Committee ranking member Ted Cruz of Texas and other Republican members appear to favor Silicon Valley venture capitalist Marc Andreessen’s anti-regulation stance, and have expressed aversion toward doing “anything on AI proactively in contrast to the Democrats.”  

Some AI experts plugged into the legislative efforts on Capitol Hill who participated in Schumer’s bipartisan AI Insight Forums would like to see Cantwell’s comprehensive bills focus on a narrow set of key issues where there has already been agreement within both major parties.

“I think if we pursue the path of bipartisanship, we should be focused on, how do we stay ahead when it comes to AI and the investments needed?” Ylli Bajraktari, CEO of the nonprofit Special Competitive Studies Project, told FedScoop.

Bajraktari said that if the bills contain too many requests for more government spending, “then you’ll have these cracks of people defecting. But if the bill maintains focus on our national security, staying ahead in innovation, and the U.S. continuing to lead, then I think that increases the chances that comprehensive bills will be bipartisan and passable.”

Paul Lekas, senior vice president for global public policy & government affairs at the Software & Information Industry Association, which represents major tech players including Adobe, Apple and Google, said it’s important that future legislative efforts follow “the bipartisan spirit” of Schumer’s AI Insight Forums. 

“It should promote and incentivize safe and trustworthy AI, mitigate potential harms to rights and safety, while allowing for continued innovation,” Lekas told FedScoop. “We encourage Congress to pass legislation establishing a nationwide standard for AI that advances public trust in the digital ecosystem, consumer confidence in AI tools, continued innovation, and U.S. competitiveness. And it should begin that effort by passing a comprehensive federal privacy bill, because AI is only as good and reliable as the data that goes into it.”

Editor’s note 1/11/2024 at 4:50 p.m.: This story was updated to reflect the rebranding of the Lincoln Network to the Foundation for American Innovation.