This study asks whether public attitudes toward autonomous weapons—and their sensitivity to contextual cues—are similar across countries. Using a preregistered survey experiment run in four countries, the authors compare how information about risk of misidentification and the identity of the target (terrorist group versus aggressive state) shifts acceptance of autonomous weapons. The paper's central contribution is a cross-national test of prior U.S.-focused experimental findings, reporting how these contextual primes operate in the United States, Brazil, Germany, and China.
preregistered survey experiment four countries risk-based primes no effect of enemy identity
Data and methods
The authors conducted a preregistered survey experiment in the United States, Brazil, Germany, and China. The design manipulated contextual information about autonomous-weapon deployment—notably a prime about the risk of target misidentification and a manipulation of enemy identity (terrorist group vs. aggressive state)—and measured respondents' expressed support for using such weapons. Exact sample sizes, fielding firms, and questionnaire text are not reported in the abstract and should be consulted in the paper or replication data for full details.
Main findings
Across the four countries, respondents showed similar reactions to the risk-based primes: information about the risk of misidentification influenced attitudes in comparable ways. By contrast, whether the target was described as a terrorist group or an aggressive state did not produce a significant effect on support. The abstract also notes that overall support levels were higher among Chinese respondents than among Americans, Brazilians, or Germans.
- Countries compared: United States, Brazil, Germany, China.
- Key contrasts: risk-of-misidentification prime (effective) versus enemy-identity prime (no significant effect).
