Who's Afraid of AI—and Why?

We live in an era where machines increasingly perform functions once restricted to humans. Automation in manufacturing has been commonplace for decades, but in the past several years self-driving cars, virtual assistants like Siri and Alexa, large language models like ChatGPT, generative artificial intelligence (AI) art programs like DALL-E, and AI military targeting systems like GOSPEL have have begun to transform our world.

These changes have brought with them economic anxiety for workers who may be displaced and ethical concerns regarding how these systems are used and the principles on which they make decisions that may mean life or death. Some have gone so far as to predict that technologies like these may lead to the destruction of humanity.

Why are so many of us so afraid of AI? Psychologists have recently begun to explore the factors that shape our attitudes towards the automation of previously human functions, uncovering individual and cultural differences in how we feel about these new developments.

Individual Differences in Fear or Acceptance of AI

In a study involving over 1,000 participants from Germany and China, researchers found that the Big Five dimensions of personality are linked to the extent to which people fear or are accepting of AI (Sindermann, et al., 2022). In the German sample, those higher in the trait of neuroticism tended to express more fear of artificial intelligence, as did Chinese participants who were higher in this trait, suggesting that people who are generally more prone to worry and anxiety were more fearful of these new technological developments. Chinese participants who were high in openness were more accepting of AI, as were those who were higher in agreeableness.

Somewhat surprisingly, among the Chinese sample (although not the German one) older participants also scored higher on AI acceptance than younger ones. And among the German sample (although not the Chinese one) women were more accepting of AI than men.

These findings suggest that personality as well as demographic characteristics may be linked to our attitudes toward artificial intelligence. Further, the fact that these linkages were often observed in one cultural group but not the other hints that cultural factors may also play a role in our feelings about GPT and its’ cousins.

Cultural Attitudes Toward AI

It turns out that the extent to which people fear AI varies substantially across societies. In a study involving 10,000 participants in 20 countries, Dong and colleagues (2023) found that fears regarding AI were generally highest in India and Saudi Arabia, and lowest in Japan.

The researchers in this work assessed concerns about AI occupying various roles, such as doctors, judges, religious leaders, and caregivers. In general, people appear most afraid of AI judges and doctors and least concerned about AI journalists and religious leaders, although there were some interesting exceptions. In Saudi Arabia, for example, concern about religious AI was particularly pronounced, as was concern about AI caregivers in Japan.

The researchers also explored factors that might be driving these effects, finding that fears about AI were linked to concerns that AI lacked the characteristics necessary to do a job in an effective and ethical fashion, i.e. warmth in the case of caregivers or fairness in the case of judges. Thus, one reason we may fear artificial intelligence taking over an increasing number of previously human roles is that we doubt that it has the psychological tendencies we feel are necessary for those jobs.

Robots and Religion

Another factor that may shape our anxiety toward AI and other forms of automation is that it may challenge people’s belief systems.

In a recent set of studies, Josh Jackson and colleagues (2023a) explored how exposure to “robot preachers” affects people's religious commitment. In one study, they took advantage of a recently automated Buddhist temple in Japan, where in one building a human preacher delivered a Buddhist sermon and in another, a robot named “Mindar” played the same role. Participants exposed to the robot preacher were less likely to donate to the temple after the sermon.

Artificial Intelligence Essential Reads

The Rise of the Techno-Energy Complex

How Leaders Can Use AI to Bring Business Objectives to Life

In another study, conducted at a Taoist temple in Singapore, the researchers arranged to have either a robot or a human deliver identical sermons on different days. Here, too, the researchers observed evidence of less religious commitment among those exposed to the robot preacher, namely a reduced likelihood of monetary donations and reduced willingness to hand out flyers for the temple. Interestingly, exposure to the robot preacher in this study also appeared to undermine the perceived credibility of the temple.

These findings also fit with another recent set of studies by Jackson’s group (2023b), including an experiment showing that learning about AI more generally reduces religious belief. Further in examining time series data from dozens of countries around the world, the researchers found that increasing exposure to automation was linked to declines in religiosity between 2006-2019 that were observed in most countries in the sample. It may be then that interacting with ChatGPT, driving alongside autonomous cars, or attending a religious service led by a robot, may cause people to question their religious worldviews.

Conclusion

Intelligent machines increasingly play a greater role in our lives, from taking us to work to creating art, writing computer code, or helping us with our homework—even leading us in worship. These changes will likely impact our psychology and appear to already be doing so. How we respond to these developments may not be uniform; it may depend on who we are and the cultures we come from.