Large Language Model Counterarguments in Older Adults: Cognitive Offloading or Susceptibility to Moral Persuasion?

Abstract

This study examined whether counterarguments generated by large language models (LLMs) influence the moral judgments of younger and older adults, and whether these effects vary by dilemma type, cognitive functioning, trust in AI, and prior LLM experience. Using the switch and footbridge trolley dilemmas, 130 participants (56 younger adults and 74 older adults) were presented with ChatGPT-generated counterarguments that opposed their initial judgments. More than 30% of participants reversed their judgments in both dilemmas (32.31% in the switch dilemma and 36.92% in the footbridge dilemma). Older adults tended to be more likely than younger adults to reverse their judgments and showed a significantly greater degree of judgment change in the switch dilemma. In the emotionally aversive footbridge dilemma, older adults with lower cognitive functioning were significantly more likely to align with the LLM-generated counterargument. General trust in AI and prior LLM experience did not predict judgment reversal, whereas lower initial confidence and higher perceived task difficulty were associated with greater susceptibility to LLM influence. These findings suggest that LLMs may support cognitive offloading but increase susceptibility among individuals with limited cognitive resources. The ecological generalizability of these findings to everyday dilemma situations remains to be examined in future research.