Microsoft President Brad Smith and the chief scientist of NVIDIA Corporation are among the witnesses slated to offer testimony on artificial intelligence Tuesday before a Senate subcommittee led by U.S. Sen. Richard Blumenthal.
Blumenthal, who chairs the Judiciary Committee’s panel on privacy and technology law, briefed reporters on his ongoing efforts to draft bipartisan artificial intelligence regulations during a Monday press conference at the state Capitol in Hartford.
“We are continuing to have hearings and conversations with the experts, industry leaders and others to form a coalition as quickly and effectively as possible behind this blueprint,” Blumenthal said of the AI regulation framework, which he released last week with Missouri Republican Sen. Josh Hawley.
The blueprint calls for the creation of an independent AI oversight body to review artificial intelligence models and require their testing before they are deployed. Other elements of the framework include provisions to ensure AI developers can be held liable for breaches in privacy or violations of human rights.
Blumenthal referred to the blueprint as a “backbone” during Monday’s press conference. It generally aims to increase transparency by requiring developers to disclose essential information about their models and include watermarks to help consumers identify AI-generated content. The proposal also calls for requiring notice when artificial systems are being used to make consequential decisions.
Tuesday’s hearing is expected to include testimony from Microsoft’s Smith along with William Dally, chief scientist and senior vice president of research at California-based tech company, NVIDIA. The subcommittee will also hear from Woodrow Hartzog, a Boston University law professor.
The subcommittee’s hearing is one of three congressional events expected to focus on artificial intelligence this week.
A House oversight committee will hold a hearing on federal agency use of AI on Wednesday. Meanwhile, Senate Majority Leader Chuck Schumer is expected to host tech executives including Facebook co-founder Mark Zuckerberg and Elon Musk, owner of X, the social media site formerly known as Twitter during a closed-door meeting Wednesday.
Blumenthal said Schumer’s attention bode well for the likelihood of passing comprehensive regulations.
“I think his taking an interest — very personally — in this subject means that we have a higher likelihood of legislation. That’s a good thing,” Blumenthal said. “What we want is not just a bill, not just hearings. We want real legislative action and the best way to do it is to continue in a bipartisan way.”
Congress will be looking to sources including Connecticut statute as it attempts to get its arms around the emergent technology, Blumenthal said. State lawmakers led by Sen. James Maroney, D-Milford, adopted a set of AI regulations during this year’s legislative session.
The bill, which passed both chambers of the legislature unanimously, was designed to scrutinize government use of artificial intelligence and algorithms in an effort to ensure automated systems are not making consequential decisions.
“Algorithms, computers, don’t stop at state borders; they don’t stop at country borders,” Maroney said during Monday’s press conference. “Making sure we have a federal approach is critical.”
Both Blumenthal and Maroney stressed that the technology could also represent enormous potential. However, Blumenthal said it was up to Congress to ensure that regulations were in place before it was too late. He said he worried about the possibility of workers in jobs ranging from writers to legal clerks at law firms being displaced by automated systems.
“We’re on the cusp of an exciting but also frightening new era portended by artificial intelligence,” Blumenthal said. “There is a panoply of peril but also tremendous promise. We want to avoid the mistake that we made with social media, which was to permit the technology to advance more quickly than the safeguards and guardrails.”
