Lawmakers Review Biden’s Order on Artificial Intelligence Development
WASHINGTON — The United States needs to innovate fast or fall behind in the international race to develop artificial intelligence, industry executives told Congress Wednesday.
The U.S. government took what one witness at the congressional hearing called “a big step” with President Joe Biden’s executive order intended to encourage development of the technology.
“It’s only one step on a longer journey,” Daniel Ho, a Stanford University Law School professor, told a House Oversight and Accountability subcommittee.
The Oct. 30 executive order was designed to set the framework for government and private industry development of artificial intelligence in the United States.
It set up a grant program for research and development that was targeted at key industries.
Some of the health care grants would fund improvements in drug development and diagnostic tools. Other grants are supposed to find solutions to climate change.
The management consulting firm McKinsey & Co. reported last week that with “the myriad increases in productivity that are likely to materialize when the technology is applied across knowledge workers’ activities — [the total economic benefit of generative AI] amounts to $6.1 trillion to $7.9 trillion annually.”
The executive order also seeks to protect against abuses of AI by assigning the National Institute of Standards and Technology, a non-regulatory agency within the Department of Commerce, to create new, non-binding standards. They would cover risks such as bank fraud and bioterrorism.
The Subcommittee on Cybersecurity, Information Technology and Government Innovation was reviewing whether Biden’s executive order is likely to achieve its goals.
Ho, who specializes in the legal issues of AI, suggested more government incentives to encourage a workforce trained to develop and use the technology.
About 40% of the doctorate-level AI workers come from foreign countries, he said. The skill shortage is acute in the federal government, where only 1% of AI workers who have a Ph.D. pursue careers in public service, Ho said.
Samuel Hammond, senior economist for the Foundation for American Innovation, said Biden’s executive order was a step in the right direction but, “My worry, though, is that it does not go far enough.”
He predicted widespread changes in industry and government workforces.
“Many companies have already downsized or plan to downsize,” he said.
Downsizing is only part of what’s coming, he said. New workforces will require more highly trained workers who know how to use AI.
Several lawmakers expressed concern about the consequences of the United States failing to lead with AI.
“China, as I’m sure you all know, is not far behind,” said Rep. Nancy Mace, R-S.C., the subcommittee’s chairwoman. “We don’t want China to catch up with us.”
Rep. Gerry Connolly, D-Va., said he was worried about abuses without better government intervention.
Armies are using AI to increase their military capabilities, he said. Faulty algorithms that direct AI programs could lead to false positives in health care, such as detecting cancerous skin lesions where there were none, an example he mentioned.
“There has to be some systematized set of standards and management,” Connolly said.
Correction: A previous version of this article referred to the National Institute of Standards and Technology as a regulatory agency. NIST is a non-regulatory agency within the U.S. Department of Commerce, as reflected in the current version.