Americans Don’t Trust AI and Want It Regulated: Poll
American companies are adopting AI technology at a rapid clip, but a majority of U.S. citizens don’t trust it and want to see its use regulated by the government, according to a new Harris Poll survey released last week. A group of tech experts that Harris Poll also surveyed showed more willingness to embrace AI, however.
The results from the survey of 2,500 adults, which were released on February 9, show that while a majority of Americans are okay with using AI for things like recommending movies or routine customer service inquiries, they are more wary about using AI for “high-value applications” such as self-driving cars, accessing government benefits, or healthcare.
Only 48% of Americans told Harris Poll that they believe AI is safe and secure, while 78% say they are very or somewhat concerned that AI can be used for malicious intent. A similar percentage (77%) said they were concerned AI could be used to generate deepfakes. Less than half (49%) said they would be okay with AI being used for a chat about “routine medical questions” and for allowing the federal government to use it to assist benefits processing.
While 71% of Americans said they agreed that AI’s primary purpose is “to assist, enhance, and empower consumers,” the poll also found that 82% of Americans would prefer that AI be regulated by the government, “to ensure adequate consumer protection.” Seventy percent also stated their desire for industry to make more investments to ensure the public is protected from AI.
“Artificial intelligence technology and frameworks could radically boost efficiency and productivity in many fields,” said Douglas Robbins, vice president of engineering and prototyping at non-profit IT and engineering firm MITRE, which commissioned Harris Poll to conduct the survey. “It can enable better, faster analysis of imagery in fields ranging from medicine to national security. And it can replace dull, dirty, and dangerous jobs.”
Harris Poll conducted a second survey for MITRE in which it asked many of the same questions to a group of 200 technology experts. In some categories, such as the potential for deepfakes to do harm, the tech experts took a harsher view of AI than the general public. However, in other categories, such as whether AI tech is “safe and secure” (66% vs 48%), whether you would be willing to use AI for everyday tasks (79% vs 48%), and whether you are comfortable with using AI to make decisions (65% vs. 37%), the tech experts took a more benevolent view of AI.
The pollsters sliced and diced the MITRE AI Survey results across demographics and unearthed some interesting trends. The poll found that Men, Democrats, younger generations, and Black/Hispanic Americans are more comfortable than their counterparts with the use of AI for some of the more advanced AI use cases, such as federal government benefits processing, online doctor bots, and autonomous, unmanned rideshare vehicles. Women, older generations, higher income, and White Americans expressed more concern for the use of AI in these categories, the survey found.
The survey of American adults was conducted online in early November via Harris On Demand omnibus. You can view the full results here.