Self-interest and preferences for the regulation of artificial intelligence

Workers have often advocated for regulation to protect themselves from labor market threats. Artificial intelligence (AI) is expected to transform many jobs in the coming years, especially those of white collar professionals. Yet we know little about the attitudes these and other workers hold or may develop toward the regulation of such technology. We argue that both self-interested and sociotropic concerns likely shape AI regulatory attitudes, and examine these arguments using a survey experiment exposing individuals to information about how technology will affect the work of people engaged in either routine, repetitive tasks or white collar, cognitive tasks. We anticipate that compared to a placebo group, individuals exposed to news about the impact of technology on jobs will favor regulation, and even more strongly when it is their own job type highlighted in the treatment. We find only modest responsiveness to the treatments on average, but observe that individuals with more knowledge of technology are strongly responsive to the AI treatments in the hypothesized manner.