Document Type
Working Paper
Publication Date
2026
Abstract
The experiment presented in this paper uses vignettes to measure the perceived likelihood of different AI futures and assess how policy preferences change across these futures. One vignette varies the degree of future AI-driven job loss in a hypothetical scenario set five years into the future, while another vignette varies the degree of future AI autonomy across the same timeframe. Participants provide their perceived likelihood that the world will resemble the scenario in the vignette, and they also express their level of support for either universal basic income (UBI) or a jobs-guarantee program (JG) if the world in five years looks like the description in the vignette. The results show that the scenarios that are perceived to be most likely have profound levels of job displacement and AI autonomy. For policy preferences, a jobs guarantee receives significantly stronger overall support than universal basic income, and support for both policies increases as job displacement – but not AI autonomy – becomes more severe. These results suggest that the public seems to be bracing for profound impacts of AI, that a jobs guarantee seems to be more politically feasible than a UBI, and that we should expect preferences for novel policies like these to be tied to levels of job loss but not necessarily to other fears related to AI advancement.
Recommended Citation
Chavanne, David, "Assessing Attitudes Toward AI Futures: Perceived Likelihood and Policy Preferences" (2026). Economics Faculty Publications. 47.
https://digitalcommons.conncoll.edu/econfacpub/47
Included in
The views expressed in this paper are solely those of the author.