Exploring User Trust in AI-Generated Summaries: A Human-Computer Interaction Perspective

چکیده

The burgeoning integration of artificial intelligence (AI) in generating textual summaries has necessitated a critical evaluation of user trust within the domain of human-computer interaction (HCI). This paper presents a comprehensive investigation into the factors influencing user trust in AI-generated summaries. By adopting a multi-disciplinary approach that synthesizes insights from cognitive psychology, computer science, and user experience design, the study delineates the complex interplay between algorithmic transparency, user interface design, and perceived summary quality.

An empirical study was conducted involving 300 participants, who interacted with AI-generated summaries across various contexts such as news articles, scientific papers, and business reports. The study employed a mixed-methods approach, combining quantitative measures of trust with qualitative feedback to capture a holistic understanding of user perceptions. The results reveal that trust in AI-generated summaries is significantly influenced by the clarity of information presentation, the perceived accuracy of content, and the transparency of the AI processes involved. Notably, the study identifies a threshold effect, whereby minor inaccuracies in summaries lead to disproportionate declines in user trust.

A theoretical model of trust was developed, encapsulating key dimensions such as usability, reliability, and transparency, which collectively inform user trust judgments. The model underscores the importance of aligning AI system affordances with user expectations to foster a sense of reliability and trustworthiness. Recommendations for designing AI systems that enhance user trust are proposed, emphasizing the need for user-centric design principles and iterative user feedback loops.

This research contributes to the HCI field by elucidating the mechanisms through which users develop trust in AI-generated content and offers actionable insights for designers and developers aiming to optimize user interaction with AI systems. The findings hold significant implications for enhancing the credibility and acceptance of AI technologies in information summarization tasks.