ACM CHI 2026 Workshop, From Generation to Simulation: Responsible Use of AI Personas in Human-Centered Design and Research, 13 April 2026, Barcelona, Spain
Tailoring writings to specific audiences increases engagement and comprehension, and LLMs with persona-based adaptation can make this process easier and scalable. However, current systems treat personas as a black-box prompt modifier: non-expert users can
specify “simplify for a beginner” or “adapt for an expert”, but cannot inspect how and what persona traits map to concrete textual edits. This opacity limits control, and prevents systematic understanding of what makes adaptations effective. We argue for decomposing
persona-driven adaptation into transparent edit operations. Rather than monolithic rewrites, we outline a taxonomy that explicitly maps persona dimensions to auditable operations. Each operation becomes a traceable composition of operators, which
specify personas, applied transformations, and design rationale, making adaptation choices visible and controllable.
Type:
Conference
City:
Barcelona
Date:
2026-04-13
Department:
Data Science
Eurecom Ref:
8704
Copyright:
© ACM, 2026. This is the author's version of the work. It is posted here by permission of ACM for your personal use. Not for redistribution. The definitive version was published in ACM CHI 2026 Workshop, From Generation to Simulation: Responsible Use of AI Personas in Human-Centered Design and Research, 13 April 2026, Barcelona, Spain
See also:
