Although artificial intelligence demonstrates high efficiency in many areas, part of the public still expresses resistance to its use. The study The Moralization of Artificial Intelligence explores whether this resistance is linked to moral beliefs rather than solely to evaluations of the risks and benefits of the technology. The authors examine whether moral attitudes toward artificial intelligence emerge as a broader social phenomenon and whether they influence users’ actual behavior.
The research consists of four interconnected studies that combine analysis of news coverage, surveys, and a longitudinal study. In the first study, 69,890 news headlines about artificial intelligence published in U.S. media between 2018 and 2024 were analyzed to determine the extent to which they used moral language. The second and third studies were conducted as surveys on representative samples of adult Americans (N=231 and N=475). Participants evaluated attitudes toward four AI applications: AI-generated art, chatbots, romantic AI partners, and legal AI systems. Researchers examined the level of opposition to the technology, moral beliefs about artificial intelligence, and opinions on whether such technologies should be banned regardless of their potential benefits.
The third study included 610 participants who evaluated seven AI applications and their non-AI counterparts, for example an AI doctor compared with a doctor via video call, as well as several other new technologies. The goal was to determine whether opposition to artificial intelligence differs from opposition to other technologies. In the fourth study, participants from previous studies took part again after 75 to 573 days. In an essay-grading task, they could use an AI grading tool that was more accurate than an average human grader, and participants also had a financial incentive for grading accuracy.
The results showed that artificial intelligence is often presented in news coverage through moral categories, at levels comparable to topics such as genetically modified organisms and vaccines. An increase in moral language in coverage of artificial intelligence was recorded after the launch of the tools DALL-E and ChatGPT in 2022. In surveys, most participants did not oppose artificial intelligence, while between 13% and 39% of respondents expressed resistance to certain applications of the technology. Among those opposed, a large proportion stated that they would remain opposed even if the benefits of the technology outweighed its risks.
The findings also show that opposition to artificial intelligence appears as a general attitude rather than merely a reaction to specific applications of the technology. In comparisons of different technologies, AI systems provoked more opposition than their non-AI alternatives. Analyses also showed that lower familiarity with artificial intelligence was associated with greater opposition.
Additionally, it was concluded that participants who showed stronger moral resistance to artificial intelligence in the first part of the study were significantly less likely to use the AI tool in the grading task later on, indicating a link between moral attitudes and the use of AI technology.
