Chuck Schumer seeks to understand AI before Congress regulates it

While talk of regulating AI might be in vogue in Washington, Senate Majority Leader Chuck Schumer wants Congress to spend some time beefing up on the particulars of AI before considering any regulation.

Speaking at an event in IBM’s Manhattan office earlier this week, the New York senator gave more details on a series of nine forums that he and a handful of colleagues plan to hold this September and October with experts from the AI community.

The goal is to understand the implications that the rise of generative AI might have, including everything from copyright and intellectual property issues and workforce changes to national security risks and doomsday scenarios, the senator said.

Schumer said he wants to hear from proponents and “skeptics—opposing views will be welcome,” and to approach the forums with a sense of “humility” in trying to understand the ins and outs of the tech.

“Humility is our watchword,” Schumer told the group of New York-based venture capitalists, startup leaders, and press at the event. “I look forward to hearing your suggestions and suggestions of the broader community as to how we can create the best safe innovation anywhere and put it into actuality.”

Schumer said that attitude will differentiate the planned forums from the grandstanding of a congressional hearing. He also named a small group of bipartisan senators who will co-lead the undertaking, including New Mexico Democrat Martin Heinrich, South Dakota Republican Mike Rounds, and Indiana Republican Todd Young.

“Frankly, congressional hearings won’t work in this regard,” Schumer said. “Having each senator or congressman get up and ask a five-minute question that may sound good back home…that’s not gonna solve this.”

Schumer told Tech Brew at the event that the forums will help Congress decide which AI regulations, recently laid out in the senator’s “Safe Innovation Framework for AI Policy,” to prioritize. “We’ll see which ones lend themselves to broader consensus dealing with innovation and safety, and which are harder. We’ll have to see,” he said.

Schumer has frequently touched on the issue of explainability—the ability to make sense of the reasoning behind black-box algorithms—as a key factor in AI regulation. But bringing transparency to the decisions of machines shaped by massive troves of data is harder than it might sound; their logic isn’t always apparent, even to their own developers.

When we asked him about the issue, Schumer acknowledged its complexity but said solving it could be critical to addressing other issues around the tech. “That’s a tough one, but it could unlock the key to a lot of other things, like intellectual property,” Schumer said. “We’re gonna have to rely on expertise like in this room because we don’t know it ourselves.”

Schumer has at least one business leader in his corner: IBM CEO Arvind Krishna said at the event that he supports AI regulation at the use-case level, and pledged his company’s support on the information-gathering process.

“You must regulate use cases, because those are what drive the benefit and the harm,” Krishna said.