Human-like Affective Cognition in Foundation Models
This work addresses the problem of assessing emotional understanding in AI for applications in human-AI interaction, though it is incremental as it builds on existing psychological theory and evaluation methods.
The paper introduced an evaluation framework to test affective cognition in foundation models, finding that models like GPT-4, Claude-3, and Gemini-1.5-Pro match or exceed human interparticipant agreement in predicting emotions from scenarios, with some conditions showing 'superhuman' performance.
Understanding emotions is fundamental to human interaction and experience. Humans easily infer emotions from situations or facial expressions, situations from emotions, and do a variety of other affective cognition. How adept is modern AI at these inferences? We introduce an evaluation framework for testing affective cognition in foundation models. Starting from psychological theory, we generate 1,280 diverse scenarios exploring relationships between appraisals, emotions, expressions, and outcomes. We evaluate the abilities of foundation models (GPT-4, Claude-3, Gemini-1.5-Pro) and humans (N = 567) across carefully selected conditions. Our results show foundation models tend to agree with human intuitions, matching or exceeding interparticipant agreement. In some conditions, models are ``superhuman'' -- they better predict modal human judgements than the average human. All models benefit from chain-of-thought reasoning. This suggests foundation models have acquired a human-like understanding of emotions and their influence on beliefs and behavior.