A new poll by the Artificial Intelligence Policy Institute (AIPI) reveals that most American voters are concerned about the risks of AI and want the federal government to regulate it. The poll, shared exclusively with Axios, shows that voters do not trust tech companies to self-regulate AI.
According to the poll, conducted by YouGov, 82% of voters do not trust tech executives to regulate AI. 56% of voters support a federal agency regulating AI, while only 14% oppose it. This shows a rare policy consensus and strong support for new federal regulation in today’s polarized electorate.
The poll also indicates that voters are more wary than excited about AI. 62% of voters are somewhat or mostly concerned about AI, while 21% are somewhat or mostly excited about it. The poll surveyed 1,001 voters from July 18-21.
Moreover, 72% of voters prefer slowing down the development of AI, compared to just 8% who want to speed it up. Three in four Democrats and Republicans alike believe that AI could eventually pose a threat to the human race. 86% of voters believe that AI could accidentally cause a catastrophic event.
The AIPI poll is not the only one that shows a preference for regulated AI. A global study by KPMG, published in February, found that 71% of people across 17 major countries believe that AI regulation is necessary. In a separate survey, 62% of British adults want to see laws and regulations around the use of AI.
However, support for AI regulation does not mean that voters in democracies trust their politicians and bureaucrats to get the rules right. The KPMG study found much lower confidence levels in government ability to regulate AI in the U.S., Japan, and the U.K. than in China, India, and Singapore.
Additionally, Americans still hold net positive views of tech, despite the decline in trust levels in recent years, according to Edelman’s annual trust barometer report.
“Americans are wary about the next stages of AI and want policymakers to step in to develop it responsibly,” said Daniel Colson, executive director of the AIPI, which favors a cautious approach to AI deployment.