Without legislation, Biden’s AI EO was broad but shallow

A long-awaited AI Executive Order was released by the Biden administration before an international AI safety summit in the UK. Without legislative support, the president can only issue so many orders, experts and stakeholders say.

The order comes as governments worldwide try to address AI’s fast-moving opportunities and risks. Long arguments and drafting processes are pushing the U.S. and EU toward action-chilling innovation.

As many companies adopt “voluntary” practices, Biden’s EO keeps them going. The president shares results, best practices, and clear guidance because they can only wave their hand.

AI risks and abuses have no legislative remedy other than tech company laws, which many argue are inadequate. Social media and de facto monopolies like Amazon and Google have been rarely regulated, but a hawkish new FTC may change that.

The prospect of a comprehensive AI law remains years away. Industry and technology have changed so quickly that any rule would be outdated by the time it was passed. What should be legislatively limited versus state law or expert agencies is unclear.

Perhaps the best approach is to create a new federal agency to regulate AI and technology, but this cannot be done by fiat. The EO instructs several AI-focused groups, such as one in the Department of Health and Human Services that handles and assesses healthcare AI-related harm reports.

Senator Mark Warner of Virginia said he was “impressed by the breadth” of the order, probably not its depth.

“I am also happy to see a number of sections that closely align with my efforts around AI safety and security and the federal government’s use of AI,” he said. However, many of these only scratch the surface, especially in health care and competition policy. Despite progress, more legislation is needed, and I will continue to work hard.

Given the state of the legislature and the upcoming contentious election, it will be a miracle if any substantive law is passed, let alone a complex and divisive AI rules bill.

NYU Stern Center for Business and Human Rights deputy director Paul Barrett acknowledged both sides.

“President Biden is sending a valuable message that certain AI systems pose immediate risks that require immediate attention. The administration is progressing, he wrote. “But today is just the beginning of a long and arduous regulatory process that will require AI companies to prove that their products are safe and effective, just as pharmaceutical, industrial chemical, and airplane manufacturers must. Without new congressional funding, the federal government may not be able to evaluate the complex training process, red-teaming, and other testing.

Tola Capital co-founder Sheila Gulati said the EO showed a “clear intention to walk the line of promoting innovation while protecting citizens.”

“We must not hinder startup agile innovation. She told that prioritizing AI explainability, risk-based approaches, and security and privacy are smart. This executive order and NIST’s standards implications suggest standards bodies will lead over legislators in the near term.

The federal government is a major customer of today’s AI and tech products, so any company that wants to keep them will need to follow the rules for the foreseeable future.

The timing seems off, according to Dorsey and Whitney partner Bob Cattanach.

The Executive Order seems to preempt Vice President Harris’s participation in a UK-hosted AI Summit, indicating White House concerns over the unregulated space and a willingness to take unilateral action to avoid delays in the EU process.

Perhaps alienate is too strong. Obviously, the UK is not the EU. That “more collaborative process” may take years, and the administration is unlikely to wait. Harris discussing the EO at the summit might have been more coherent and ally-like. On November 1, you can stream her remarks, which will likely call for international AI regulation, with the US modestly leading.