Maritza Ramirez Mills

(This poster summarizes a paper to be published in the ACM AI Leadership Summit 2026 proceedings) This paper proposes that emotion understanding in large language models (LLMs) is asymmetric to human affective cognition and therefore faces an inevitable performance ceiling when applied to areas which require sustained interpersonal relating, such as mental health care. While LLMs share in the correlative, grammatical and inductive aspects of emotion understanding, they differ in that they do not share in the same non-linguistic, interoceptive sources of information that humans do. This makes the lexical representations they build asymmetric and incomplete when compared with humans. Therefore, any apparent emotion understanding in LLMs is the result of the successful elicitation of emotion understanding in human interlocutors across a dynamic representational medium. The knowledge itself rests in the source and destination, rather than the intermediate representation generated by LLMs. This fundamental asymmetry is not solved by scaling toward multi-modal variants of third personal data. Research directions which prioritize emotion self-understanding in humans, along with the facilitation of human-human interaction, will advance Emotion AI further than multi-modal integration alone.

Since inception, AI research has been inspired by the idea of building machines that think and behave like humans. Or put another way, machines that think and act rationally. This paradigm has its historical foundations in the assumption that what makes humans different from other forms of life is our capacity for self-reflective, rational thought. This assumption has caused research into artificial intelligence and human intelligence more generally to undervalue the role of emotions in generalized problem-solving. The very definition of rationality has, for much of Western history, discounted emotions from its conception. Where emotions are investigated, they are often treated as irrational, immaterial, or illusory aspects of human experience. This has been an unfortunate mistake.

There has been an explosion of interest in understanding LLM creativity. A review of current literature suggests that exactly what creativity is and how to measure it remains unclear, with further disagreement regarding how to apply existing research in human creativity to artificial intelligence. This poster provides an overview of recent approaches used in computing and recommends future directions by integrating findings from philosophy and psychology. Measures in scope are those with the goal of assessing LLM response creativity and what factors influence more or less apparent creativity in response output.