Bloomberg Law
May 19, 2023, 9:15 AM

AI Threats Confront Eager Congress Grappling With Learning Curve

Oma Seddiq

Leaders of artificial intelligence companies are making an unusual request to Congress as their industry advances: establish safeguards to mitigate threats their technology poses.

The “stakes are too high for Congress not to act at this moment,” said Ryan Hagemann, co-director of IBM Policy Lab from International Business Machines Corp,which has been calling for risk-based regulation for AI for more than three years.

Tools such as OpenAI Inc.’s ChatGPT that collect data from users have raised privacy concerns. New language models have the ability to write songs, produce visuals, and mimic real-life artists without their consent. Human bias and discrimination could permeate AI systems and negatively affect minorities. The rise in deepfakes and voice cloning could pose risks to democracy, especially regarding misinformation in elections. Bad actors could exploit AI to spam and manipulate users.

See also: NYC Public Schools Allow ChatGPT Use After ‘Knee-Jerk’ Ban

Though Democrats and Republicans have displayed unity over the need to regulate AI, history isn’t on Congress’ side when it comes to reining in the tech industry. Lawmakers failed to significantly move the needle on social media regulations and now worry about bungling the response to AI.

“I’m skeptical because I know how Congress responds. It takes time. Too much time,” Dick Durbin (Ill.), the No. 2 Democrat in the Senate, said. “And we’re reluctant to gather the resources for something this challenging. So, it will be amazing if we do something, and we should.”

‘Degree of Humility’

Congress has begun taking steps to address AI. The Senate Judiciary’s privacy, technology, and law subcommittee heard testimony this week on AI’s dangers from OpenAI CEO Sam Altman and IBM Chief Privacy and Trust Officer Christina Montgomery. The hearing was the first of many to better understand the benefits and harms of the emerging technology.

Read More: OpenAI, IBM Urge Senate to Act on AI Regulation After Past Tech Failures

Senate Majority Leader Schumer (D-N.Y.) met Wednesday with Sens. Martin Heinrich (D-N.M.), Mike Rounds (R-S.D.), and Todd Young (R-Ind.) about their emerging bipartisan group focused on comprehensive AI legislation, according to a source familiar with the meeting.

“Our group agreed that because AI technology is developing so quickly, Congress has to move fast,” Schumer said on the Senate floor Thursday.

But discussions around AI are a far cry from moving forward on sweeping legislation.

“The idea that we need to put some guardrails in place is a good one. But what does that mean? I don’t know,” said Josh Hawley (R-Mo.), ranking member of the Senate Judiciary panel that hosted the hearing this week. “Nobody knows, right?”

Sen. Richard Blumenthal (D-Conn.), chair of the subcommittee, acknowledged regulating AI is a complex task and Congress doesn’t have the expertise.

Sen. Richard Blumenthal (D-Conn.) asks questions before the Senate Judiciary Subcommittee on Privacy, Technology, and the Law on May 16, 2023 in Washington, D.C.
Photographer: Win McNamee/Getty Images

“That degree of humility is required,” said Blumenthal, who opened the hearing with an AI-generated recording that sounded like his voice.

Advocates for AI regulation say they are enthused by lawmakers’ recent focus on the technology, though also recognize that a learning curve exists.

“That, I think, is going to be the question: How quickly can they feel like they have enough competency in this space to draft a bill that actually gets at some of the harms that they unearthed?” said Jordan Shapiro, director of the Progressive Policy Institute’s Innovation Frontier Project.

See also: AI Leader Sees a Rare Chance in Congress for Wins on Regulation

Policy Ideas

Lawmakers, company officials, and policy experts agree Congress can put at least some measures in place to target the technology responsibly and constructively, such as protecting artists’ rights.

“It’s not sustainable for us to live in a world where big AI companies scrape all this protected content and displace all of these creative industries,” said Michael Huppe, president and CEO of SoundExchange, a collective-rights management company for musicians. “I’m optimistic that we’ll find a way through this because we have to.”

Suggestions at the Judiciary hearing included creating a federal agency to issue AI rules or empowering individuals harmed by AI with the right to sue the companies behind it.

Earlier: ChatGPT App Comes to iPhone, With Android Version in Works (1)

Schumer last month announced he’s been working on a regulatory framework that would require independent reviews and testing of AI technologies before they go public.

Yet Congress should exercise caution when passing regulations to ensure they bring about change without stifling technological progress, policy experts warn.

“Regulating the entire technology is going to hurt innovation,” Shapiro said. “It’s going to be much easier and much more impactful to define specific harms and then try to regulate for those harms.”

Another idea is requiring disclosures on AI-generated content to promote transparency to users and prevent them from being deceived. Democratic Sens. Amy Klobuchar (Minn.), Cory Booker (N.J.), and Michael Bennet (Colo.) earlier this week unveiled legislation that would require a disclaimer on political ads for federal campaigns that use AI.

“The election is quite top of mind and we know the effects of misinformation on it,” Shapiro said. “We want to have an open, truthful, cordial election. And I think that is an area that perhaps is almost a low-hanging fruit that Congress could address.”

As part of the growing momentum, Congress spotlighted AI in two other panel hearings held this week.The House and Senate each have bipartisan AI caucuses. Rep. Don Beyer (D-Va.), vice chair of the House AI Caucus, is pursuing a master’s degree in machine learning, partly to better inform Congress’ work. The Biden administration last fall released a Bill of Rights that outlined best practices for AI developers and users.

Read More: AI Companies Should Pay to License Work, Artists Tell Lawmakers

For his part, OpenAI’s chief Altman struck an optimistic tone on whether Congress will act.

“I hope that they will. They seem sincere, and if they don’t — then the industry does what we think we should do anyway,” he told reporters after this week’s hearing.

— With assistance from Zach C. Cohen.

To contact the reporter on this story: Oma Seddiq at oseddiq@bloombergindustry.com

To contact the editors responsible for this story: Anna Yukhananov at ayukhananov@bloombergindustry.com; Robin Meszoly at rmeszoly@bgov.com

Learn more about Bloomberg Law or Log In to keep reading:

Learn About Bloomberg Law

AI-powered legal analytics, workflow tools and premium legal & business news.

Already a subscriber?

Log in to keep reading or access research tools.