Schumer outlines plan for how Senate will regulate AI
Views: 1541
2023-06-22 01:58
Senate Majority Leader Chuck Schumer announced a broad, open-ended plan for regulating artificial intelligence on Wednesday, describing AI as an unprecedented challenge for Congress that effectively has policymakers "starting from scratch."

Senate Majority Leader Chuck Schumer announced a broad, open-ended plan for regulating artificial intelligence on Wednesday, describing AI as an unprecedented challenge for Congress that effectively has policymakers "starting from scratch."

The plan, Schumer said at a speech in Washington, will begin with at least nine panels to identify and discuss the hardest questions that regulations on AI will have to answer, including how to protect workers, national security and copyright and to defend against "doomsday scenarios." The panels will be composed of experts from industry, academia and civil society, with the first sessions taking place in September, Schumer said.

The Senate will then turn to committee chairs and other vocal lawmakers on AI legislation to develop bills reflecting the panel discussions, Schumer added, arguing that the resulting US solution could leapfrog existing regulatory proposals from around the world.

"If we can put this together in a very serious way, I think the rest of the world will follow and we can set the direction of how we ought to go in AI, because I don't think any of the existing proposals have captured that imagination," Schumer said, reflecting on other recent proposals such as the European Union's draft AI Act, which last week was approved by the European Parliament.

The speech represents Schumer's most definitive remarks to date on a problem that has dogged Congress for months amid the wide embrace of tools such as ChatGPT: How to catch up, or get ahead, on policymaking for a technology that is already in the hands of millions of people and evolving rapidly.

In the wake of ChatGPT's viral success, Silicon Valley has raced to develop and deploy a new crop of generative AI tools that can produce images and writing almost instantly, with the potential to change how people work, shop and interact with each other. But these same tools have also raised concerns for their potential to make factual errors, spread misinformation and perpetuate biases, among other issues.

In contrast to the fast pace of AI advancements, Schumer has stressed the importance of a deliberate approach, focusing on getting lawmakers acquainted with the basic facts of the technology and the issues it raises before seeking to legislate. He and three other colleagues began last week by convening the first in a series of closed-door briefings on AI for senators that is expected to run through the summer.

In his remarks Wednesday, Schumer appeared to acknowledge criticism of his pace.

"I know many of you have spent months calling on us to act," he said. "I hear you. I hear you loud and clear."

But he described AI as a novel issue for which Congress lacks a guide.

"It's not like labor, or healthcare, or defense, where Congress has had a long history we can work off of," he said. "Experts aren't even sure which questions policymakers should be asking. In many ways, we're starting from scratch."

Schumer described his plan as laying "a foundation for AI policy" that will do "years of work in a matter of months."

To guide that process, Schumer expanded on a set of principles he first announced in April. Formally unveiling the framework on Wednesday, Schumer said any legislation on AI should be geared toward facilitating innovation before addressing risks to national security or democratic governance.

"Innovation first," Schumer said, "but with security, accountability, [democratic] foundations and explainability."

The last two pillars of his framework, Schumer said, may be among the most important, as unrestricted artificial intelligence could undermine electoral processes or make it impossible to critically evaluate an AI's claims.

Schumer's remarks were restrained in calling for any specific proposals. At one point, he acknowledged that a consensus may even emerge that recommends against major government intervention on the technology.

But he was clear on one point: "We do — we do — need to require companies to develop a system where in simple and understandable terms users understand why the system produced a particular answer, and where that answer came from."

The Senate may still be a long way off from unveiling any comprehensive proposal, however. Schumer predicted that the process is likely to take longer than weeks but shorter than years.

"Months would be the proper timeline," he said.

Tags ai schumer epus one epus scitech plan eppersons