An Adversarial Dual-Agent Critical Framework for Intelligent Evaluation and Optimization of Human–Computer Interaction Design
Abstract
Critical thinking is a core practice in the field of human-computer interaction and design, aiming to enhance the quality of interaction solutions in multiple dimensions such as experience, technology, and ethics through systematic review. Existing research lacks collaborative deduction and in-depth demonstration of HCI schemes at the behavioral logic, engineering implementation and comprehensive risk levels. This leads to the difficulty in improving love, with feedback remaining superficial and fragmented, making it hard to support high-quality innovation and decision-making in a complex and dynamic interactive context. This paper proposes an adversarial design critical framework based on dual agents. This framework consists of two adversarial agents: (1) Experience Optimization Agent: This agent takes user experience as the core evaluation dimension and quantitatively analyzes the intuitiveness, operational efficiency, and user emotional feedback of the design plan based on interaction design principles, cognitive psychology models, and input user expectations. (2) Constraint Verification Agent: This agent takes technical feasibility and actual conditions as the evaluation dimensions, and based on implementation cost, performance indicators, multi-terminal adaptation requirements and basic design specifications, identifies technical implementation risks, performance defects and compliance issues existing in the design scheme. The results of user experiments show that, compared with traditional schemes, the framework constructed in this paper demonstrates outstanding performance in human-computer interaction tasks, significantly enhancing the comprehensive adoption rate of the generated suggestions, and showing efficient early warning capabilities for key design issues such as logical contradictions and imbalance of resource benefits. The collaborative evaluation mechanism effectively reduces the cognitive load generated when dealing with multi-dimensional feedback, enabling designers to focus more on core design decisions. In complex design contexts such as internationalization and cross-cultural, this framework also demonstrates superior adaptability and strategy generation potential, which is conducive to promoting the evolution of design assistance tools towards an intelligent collaboration paradigm with higher-level cognitive support capabilities.
Keywords: Critical Thinking, Human-computer Interaction, Design, Dual Agents
DOI: 10.54941/ahfe1007512
Cite this paper
More from this volume
- Brain-Computer Interface versus Brain-Computer Interaction
- Human–AI Interaction as a Catalyst for Interdisciplinary Co-Creation: Exploring Prompt-Driven Visualization in Design Education
- Context-aware LLMs for healthcare requirements engineering
- Understanding the Needs and Challenges of Developing Robot Teleoperation Applications using Mixed Reality Headsets
- Daughter-Led Intergenerational Collaboration: Human-Computer Interaction in APP-Based IUD Removal Support for Midlife Women
- The Effect of the Degree of Multimodal Information Explanation by AI Streamers on Consumers’ Purchase Intention: The Moderating Role of Product Type
- Refining Research Questions for AI-Assisted Knowledge Retrieval in Interior Design: An Exploratory Study of Expert Judgment
- Performance Trust in AI Reduces Cognitive Workload: Evidence from Structural Equation Modeling and Item-Level Analysis
- The Impact of Direct and Third-Party Control: A Comparison of the Usage of AI Advice in Hiring Decisions
- User Perceptions of Response Inconsistency and Trust in AI-Assisted Learning
- Designing a Rhythmic AR Interaction for Auditory-Oriented Heritage: A Preliminary Case Study at Guqintai
- Feedback-Driven Adaptive AR Assistance for Intralogistics: Design and Initial Evaluation


AHFE Open Access