ABSTRACT: As AI systems increasingly make decisions without visible interfaces, the user experience is shaped less by interaction and more by outcome. This shift brings new challenges—especially when adaptive systems begin to misinterpret their purpose or apply the wrong logic in high-stakes environments like finance, public services, or healthcare.
This talk introduces the TAR Framework—Trust, Alignment, Recourse—a governance model developed by me and my team at Phase 5 to address what we’ve identified as AI Identity Drift. We’ll explore how AI can quietly lose alignment with its original role, and how UX and design teams can ensure these systems remain explainable, grounded, and challengeable.
The framework is grounded in UX and CX principles, but it extends into governance, compliance, and institutional trust—offering a practical, experience-centered approach to designing and overseeing AI systems that evolve over time.
BIO: Arnie Guha, Ph.D. is a Partner and Head of the User Experience Strategy and Design practice at Phase 5, an insights consultancy that helps organizations Make Bold Decisions across Brand, CX, UX, and Innovation. At Phase 5, Arnie leads multidisciplinary teams focused on the human impact of emerging technologies, including the development of the TAR Framework (Trust, Alignment, Recourse) to ensure AI systems remain intelligible, aligned, and challengeable. Arnie also serves as Chair of the Advisory Board at Green College (UBC), President of The National Club in Toronto, and Chair of the Aird Gallery. He is the creative visionary behind Acid4Yuppies, a multidisciplinary art project exploring perception, experience, and consciousness through lightboxes, silk accessories, interior design, and immersive installations.