Moral Judgment in Groups of Human or AI Agents

Dr. Anita Keshmirian, our Head of Data Science & Assistant Professor in Psychology at Forward College, was invited to present her research at Aalto University’s in the Society & Networks Seminar. Her talk, ‘Moral Judgment in Groups of Human or AI Agents,’ examines how social interaction shapes moral judgments in humans and whether similar patterns emerge when large language models (LLMs) deliberate in small groups. The research finds that group deliberation leads to more utilitarian judgments in both humans and LLMs, although the underlying mechanisms appear to differ: in humans, the effect is accompanied by reduced anxiety, while in LLMs, it is better explained by changes in how norm violations and beneficiaries are weighted. The talk also discusses the implications of these findings for AI alignment, multi-agent system design and artificial moral reasoning.