Agentic UX: A Framework for Trust and Transparency in AI Autonomy
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
AI has progressed rapidly, ranging from conventional "Command Response" models to self-governing "Agentic" systems capable of automatically defining goals and achieving them through a series of steps.Although the algorithmic reasoning process of these agents has significantly improved, the User Experience associated with AI remains a bottlenecking point.Since current interfaces exist for static software, it causes users to feel effects of "Automation Bias" when the software works automatically as well as "Transparency Gap" when the software works automatically.This paper presents a new theoretical framework called "Agentic UX," which is intended to enable smooth human-AI cooperation.Three fundamental pillars are introduced by the framework: (1) Confidence-Based Handoffs, a protocol that identifies high-risk situations where control must be returned to humans; (2) Proactive Transparency, which shows the AI's internal chain of thought to ensure explainability; and (3) Adjustable Autonomy, a mechanism that enables users to adjust the AI's independence based on task complexity.By changing the design paradigm from "Interfacing with a Tool" to "Collaborating with a Partner," this study offers a guide for creating autonomous systems that are transparent, reliable, and user-focused.The study comes to the conclusion that the clarity of the interface that controls the agent's behaviour will determine how widely AI is adopted in the future rather than the agent's strength.
Fields of Study & Subjects
Human-Automation Interaction and Safety
Usability and User Interface Design
Innovative Human-Technology Interaction
Publication details
- Journal / source
- International Research Journal of Modernization in Engineering Technology and Science
- Publication date
- 1 April 2026
- Language
- EN