• by Ray Schultz , Columnist, April 9, 2024

American voters are suspicious of AI, and many favor government restrictions on it. Many fear the spread of disinformation, and most say publishers should be compensated for use of their content in training AI tools, according to a study by the News/Media Alliance. 

Overall, 31% are very uncomfortable with AI and 35% are somewhat so. The most uneasy individuals are seniors (81%), rural voters (77), women (71%) and conservatives (71%). 

At the same time, 57% support compensating publishers for use of their material in training. This feeling is strongest among college graduates (50%), Democrats (52%) and voters who get their news through group messaging apps (75%), websites (66%), and daily (64%) or weekly (63%) newspapers.

“This should call into further question users’ trust and reliance on products that were built on stolen property,” says Danielle Coffey, president & CEO of News/Media Alliance. “The notion that massive amounts of content cannot possibly be licensed flies in the face of current large scale licensing of content for entertainment, R&D, financial, governmental, academic, medical, and other endless uses.”

In addition, 72% feel that limitations should be placed on AI, including 36% who strongly support this idea and 36% who somewhat agree with it.  

This shows a desire for government action not seen when other new technologies were introduced, the study says. 

The respondents are also very concerned about these suspicions:

Artificial intelligence (AI) fuels the spread of misinformation by making it easy to generate and post fake news articles and misleading content online—66%

Artificial intelligent compromises the integrity of our elections by making it easier to spread partisan and biased information—60%

News generated by artificial intelligence jeopardizes Americans’ ability to have trustworthy and transparent sources of information—59%

The News/Media Alliance surveyed 1,800 registered voters from February 3-11.