Abhorrence and Autonomy: On the Place of Public Attitudes in the Debate on Militarised Artificial Intelligence

Abstract Advocates of banning autonomous weapons systems invoke the Martens Clause, arguing that widespread public opposition constitutes “dictates of public conscience” sufficient to ground prohibition absent explicit treaty language. This claim rests on polling reporting stable opposition to autonomous weapons, interpreted as a categorical moral conviction against delegating life-and-death decisions to machines. This article argues that this abhorrence-based case fails to justify, with current evidence, a blanket prohibition. Drawing on experimental research in moral psychology and studies of public attitudes toward autonomous weapons, I show that these intuitions fail to meet the necessary epistemic standards, due to framing issues, differences across demographic categories, and plausibly due to structural concerns regarding how advocacy groups are incentivised to shape public opinions rather than simply measure them. This critique neither establishes moral acceptability of autonomous weapons nor rejects intuition-based arguments in principle; rather, it clarifies epistemic limits of current appeals to public conscience and their inability to meet the threshold of “crystallised moral consensus” that the Martens Clause requires. The article concludes by proposing systematic experimental methods capable of identifying stable moral opposition to specific high-risk configurations of military autonomy, supporting discriminating regulation rather than categorical prohibition.