Keito Inoshita

The guiding question of his research is consistent: how can AI engage with human emotion without diminishing its richness? Emotion is ambiguous and polysemous, yet it shapes human judgment and relationships in every corner of society. Treating that ambiguity as an object of computation, rather than discarding it, is the starting point of his work.

To this end, he conducts coherent research spanning the generation, recognition, and understanding of human emotion with AI, human-AI interaction, and the ethics and philosophy of affective AI. Recognition becomes a technique for estimating ambiguity as a distribution; generation and expression form the interface where humans and AI converse through emotion; understanding reopens the question of how people experience emotion in the first place; and ethics and philosophy structurally reconsider the social positioning of affective AI.

Beyond academic work, through partnerships with diverse companies he incorporates field-level insight from the sites where affective AI is deployed and reaches society into the foundation of his research. The aim is to draw the contours of AI endowed with an EQ that surpasses humans, through a back-and-forth between technology and the humanities, and between implementation and theory.