Schatz, Kennedy File Legislation Requiring Public AI Disclosures
Sens. Brian Schatz, D-Hawaii, and John Kennedy, R-La., introduced legislation Tuesday that would require generative AI developers to disclose when their content is AI-generated. Schatz and several senators told us in interviews they believe AI can be regulated with or without the passage of a federal privacy law.
“Anything made with generative AI should be disclosed,” Schatz told us Tuesday. The bill would implement a “simple, watermark requirement for visual stuff and a verbal disclosure for audio files.” Other AI policy issues require “further investigation and nuance,” he said. “This is not one of them. People are entitled to know” when content is created by AI. Asked if the technology can be regulated without a federal privacy law, Schatz said, “Yes.”
The AI Labeling Act would require developers to include a “clear and conspicuous disclosure identifying AI-generated content and AI chatbots.” Developers and third-party licensees would have to take “reasonable steps to prevent systematic publication of content without disclosures.” The bill includes a recommendation that social media platforms form a working group to create nonbinding technical standards for identifying AI-generated content.
The bill “would set an AI-based standard to protect U.S. consumers by telling them whether what they’re reading, seeing or hearing is the product of AI, and that’s clarity that people desperately need,” Kennedy said in a statement.
House Commerce Committee Chair Cathy McMorris Rodgers, R-Wash., told us last week that her committee continues to work toward reintroducing bipartisan privacy legislation. She and party leaders on the committee said last week they plan to expand their bill from 2022 to address AI technology (see 2310180051). Asked if the committee is waiting to get Senate buy-in before reintroducing the bill, Rodgers said, “We are in discussions with the Senate.” Senate Consumer Protection Subcommittee Chairman Gus Bilirakis, R-Fla., told us last week: “We want to build consensus, and obviously we want the Senate on board.”
Sens. Richard Blumenthal, D-Conn.; Martin Heinrich, D-N.M.; John Hickenlooper, D-Colo.; and Mark Warner, D-Va., Senate Intelligence Committee chairman, told us in separate interviews Congress can move forward with AI regulations, regardless of what happens with privacy legislation. “This is the issue in front of us, and I think we have enough consensus to begin regulating AI,” said Heinrich.
Blumenthal pointed to his AI framework with Senate Privacy Subcommittee ranking member Josh Hawley, R-Mo., as a potential path forward. That will be the focal point as the subcommittee nails down a date for another AI hearing in November or December, he said.
“I think ultimately we will have to have a privacy law,” said Hickenlooper. “It certainly will make how we approach AI laws more direct if we have a privacy law first.” But AI is moving so “rapidly” that Congress needs to move forward with AI policies. Hickenlooper told us he planned to attend Tuesday’s all-Senate AI forum hosted by Senate Majority Leader Chuck Schumer, D-N.Y.
Tuesday’s Senate forum attendees included AFL-CIO Technology Institute Director Amanda Ballantyne, Stripe CEO Patrick Collison, Carnegie Mellon University professor Jodi Forlizzi, NAACP CEO Derrick Johnson, Center for Democracy and Technology CEO Alexandra Reeve Givens and SeedAI President Austin Carson, according to Schumer’s office. Congress should focus on increasing transparency and security and reducing automated bias, Schumer said on the Senate floor Tuesday.