Nearly two-thirds of Americans want the U.S to regulate the development and use of artificial intelligence in the next year or sooner — with half saying that regulation should have begun yesterday, according to a Morning Consult poll. Another 13% say that regulation should start in the next year.
"You can thread this together," Austin Carson, founder of new nonprofit group SeedAI and former government relations lead for Nvidia, said in an email. "Half or more Americans want to address all of these things, split pretty evenly along ideological lines."
The poll, which SeedAI commissioned, backs up earlier findings that while U.S. adults support investment in the development of AI, they want clear rules around that development. Almost 70% of adults support more work in developing AI, but majorities from both parties — 67% of Democrats and 56% of Republicans — said that they'd want to see regulation sooner if they knew companies were already testing for bias and other issues.
Carson's group, which officially launches Friday, aims to make sure underrepresented groups and regions can learn about AI, train for jobs in the sector and benefit from the technology. He said he wanted to see the U.S. providing "resources for a diverse range of communities" that would be available for research, development, training and testing on trustworthiness. The formal launch of Carson's group will include Lynne Parker, who is director of the National Artificial Intelligence Initiative Office, as well as members of Congress as speakers.
The poll found that approximately eight in 10 respondents think the federal government should probably or definitely try to make sure AI is trustworthy and safe, with an even greater share supporting a role for state governments or efforts by the companies themselves. On the topic of limiting bias, 65% said the federal government should probably or definitely work on the issue, with 69% saying private companies should.
The itch for regulation is unlikely to get a scratch from the feds anytime soon. Congress has repeatedly failed to pass significant laws governing much more mature aspects of technology, such as the ongoing push for a nationwide privacy law. Still, lawmakers have proposed AI-regulating measures in the past, and a Senate-passed bill that aims to boost the U.S. against China would set up a scholarship to "recruit and train artificial intelligence professionals to lead and support the application of" AI to government activities.
Industry also wants Congress to act. In June, the BSA, a trade group for many companies that produce AI systems, began advocating for Congress to enact a law requiring firms to assess and minimize bias for "high-risk" uses of AI. BSA said at the time it believed the process could take years and added its hope that the proposal would replace farther-reaching legislation.
Regulators elsewhere, however, are also taking notice. Earlier this year, Europe suggested banning systems that pose a serious risk to health and human rights, as well as strict rules for an array of "high-risk" AI.
The poll suggested some specific areas where the public might like to see regulators take action. More than 70% of the people who responded said they were at least somewhat worried about AI being vulnerable to hacking or used to spread misinformation. The majority of respondents also expressed some worry about bias, the decline of human autonomy in daily tasks, job losses, over-reliance on AI in policing or military applications, and AI systems "becoming uncontrollable."
The results suggested the public sees benefits to AI as well. Strong majorities favored investments in teaching kids about using AI and training workers for jobs in the field, particularly the aim of making the U.S. "a global leader" in both areas. Nearly 80% said U.S. investment in AI research and development for military and defense applications was at least somewhat important, alongside 75% who said so for health care.
Interestingly, while Baby Boomers had higher levels of concern about AI, the youngest adults in the survey, from Gen Z (born after 1997), were less likely to say the U.S. should have begun its regulation already, or to be interested in learning about current or future uses of AI.
"As soon as you drop to digital natives, the game changes," Carson said. He suggested that younger adults' attitudes could evolve, but are probably "a more accurate barometer for the future" than millennials and older generations.
The poll interviewed 2,200 adults online, and full survey results had a margin of error of plus or minus 2 percentage points.