The United States Congress seeks to balance innovation and control in the development of artificial intelligence

Last month, House leaders took a major step toward drafting an action plan on the issue artificial intelligence With the creation of a new bipartisan task force, which will issue recommendations on how Congress can boost AI innovation while preserving the tools under control.

However, lawmakers leading the group said Washington Post In a joint interview, “implementing the full response is likely to be a tall order,” as they consider technology's massive impact on elections, national security, the economy, and more.

California Republican Rep. Jay Obernault, chosen by House leaders to chair the group, pointed to efforts by Europe to agree on a “comprehensive” law on artificial intelligence as a cautionary example. “If you look at European attempts to create a common law to regulate artificial intelligence, you will see some of their fallacies,” said Obernault, one of the few lawmakers with computer science experience. “They've had to rewrite this bill several times as the face of AI has changed.”

“We cannot imagine a 5,000-page bill that addresses 57 issues and eliminates artificial intelligence,” said Democratic Rep. Ted Lieu of California, co-chair of the task force. “It will be a multi-year process, and there will be a variety of different bills that will try to address different aspects of AI.”

The working group is scheduled to issue a report by the end of the year, but that does not preclude more immediate legislative action on specific issues, Obernault and Liu said.

Like Senate Majority Leader Charles E. Schumer, Democrat Charles E. Schumer, Obernault pointed to the risks that AI-generated content poses to elections as an area with the potential for quick action. “There must be broad bipartisan agreement not to allow anyone to impersonate a candidate using artificial intelligence“So we will study what we can do to tighten the regulations and try to avoid it,” he said.

Liu supported the idea and suggested criminal and civil “improvements” to increase fines or prison sentences for certain AI crimes. “One way to increase deterrence is to say that if AI is used to impersonate a voice that defrauds someone, that will increase the penalty that can be imposed,” he said.

Obernault expressed “hope” that Congress will prioritize the Create Artificial Intelligence Act, which aims to launch the National Artificial Intelligence Research Resources (NAIRR). In January, The White House has launched a pilot version of the center, which is scheduled to last for two years.

In the Senate, Schumer was criticized by some colleagues for keeping a series of AI “information forums” closed to the public. (In response, he noted that the House has held numerous public hearings on AI over the years.)

In the House of Representatives, Obernault and Liu said they plan to hold public and private sessions to delve deeper into the many aspects of artificial intelligence. “We want to hold open meetings in a traditional hearing format to make sure we are transparent with the public,” Obernault said. “But we will also have some closed meetings… because it is very important to me that everyone feels comfortable asking questions that may seem ignorant.”

Although the bipartisan AI task force has not yet come up with any proposals or legislative text, it predicted in June that the Senate would act within “months” rather than “years.” For their part, House leaders did not launch the task force until nearly a year after Schumer unveiled his plans, which It raised concerns among some members about the Council's absence from the discussion.

Obernault and Leo rejected these suggestions. “We will make progress over the next few years, and we can do so because there are short-, medium- and long-term damages that need to be mitigated,” Obernault said. “I don't think this is at all inconsistent with what the Senate is doing.”

He added that their offices have had “informal contact over the past year” with Schumer's task force leaders, but said “they are very aware that we want to work with them and I think they are very open to working with us.” “We're just getting started,” Leo agreed.

(c) 2024, The Washington Post