Psychological and Brain Responses to Artificial Intelligence's Violation of Community Ethics.

Human moral reactions to artificial intelligence (AI) agents' behavior constitute an important aspect of modern-day human-AI relationships. Although previous studies have mainly focused on autonomy ethics, this study investigates how individuals judge AI agents' violations of community ethics (including betrayals and subversions) compared with human violations. Participants' behavioral responses, event-related potentials (ERPs), and individual differences were assessed. Behavioral findings reveal that participants rated AI agents' community-violating actions less morally negative than human transgressions, possibly because AI agents are commonly perceived as having less agency than human adults. The ERP N1 component showed the same pattern with moral rating scores, indicating the modulation effect of human-AI differences on initial moral intuitions. Moreover, the level of social withdrawal correlated with a smaller N1 in the human condition but not in the AI condition. The N2 and P2 components were sensitive to the difference between the loyalty/betrayal and authority/subversion domains but not human/AI differences. Individual levels of moral sense and autistic traits also influenced behavioral data, especially on the loyalty/betrayal domain. In our opinion, these findings offer insights for predicting moral responses to AI agents and guiding ethical AI development aligned with human moral values.

Dissociated modulations of intranasal vasopressin on prosocial learning between reward-seeking and punishment-avoidance.

BACKGROUND: As an integral ingredient of human sociality, prosocial behavior requires learning what acts can benefit or harm others. However, it remains unknown how individuals adjust prosocial learning to avoid punishment or to pursue reward. Given that arginine vasopressin (AVP) is a neuropeptide that has been involved in modulating various social behaviors in mammals, it could be a crucial neurochemical facilitator that supports prosocial learning. METHODS: In 50 placebo controls and 54 participants with AVP administration, we examined the modulation of AVP on the prosocial learning characterized by reward and punishment framework, as well as its underlying neurocomputational mechanisms combining computational modeling, event-related potentials and oscillations. RESULTS: We found a self-bias that individuals learn to avoid punishment asymmetrically more severely than reward-seeking. Importantly, AVP increased behavioral performances and learning rates when making decisions to avoid losses for others and to obtain gains for self. These behavioral effects were underpinned by larger responses of stimulus-preceding negativity (SPN) to anticipation, as well as higher punishment-related feedback-related negativity (FRN) for prosocial learning and reward-related P300 for proself benefits, while FRN and P300 neural processes were integrated into theta (4-7 Hz) oscillation at the outcome evaluation stage. CONCLUSIONS: These results suggest that AVP context-dependently up-regulates altruism for concerning others' losses and reward-seeking for self-oriented benefits. Our findings provide insight into the selectively modulatory roles of AVP in prosocial behaviors depending on learning contexts between proself reward-seeking and prosocial punishment-avoidance.