Stories about HCXAI
1 related stories
Not All Explanations Are Sought: Information-Seeking Psychology for Human-Centered XAI
AI InsightThis position paper proposes integrating the psychology of information seeking into human-centered explainable AI (HCXAI). Compared to prior technical paths focused on generating more detailed explanations, this adds a new dimension: evaluating the cognitive mechanisms of whether users are willing to consume explanations. It notes users decide whether to seek explanations based on instrumental, hedonic, and cognitive expected utilities, a process influenced by six cognitive biases like illusion of control and automation bias. This implies HCXAI system design must shift from unidirectional output to actively intervening in users' explanation avoidance or over-seeking behaviors.Key TakeawayHCXAI focus shifts from generating explanations to analyzing users' consumption motives.Why It MattersProvides a predictable theoretical framework for the interaction layer design of XAI systems, directly impacting future explanation presentation strategies.Who's Affected- AI ResearchersExpands HCXAI boundaries, prompting inclusion of cognitive psychology in AI explanation efficacy metrics.
- DevelopersMust build dynamic explanation strategies addressing cognitive biases rather than merely stacking information.
What's NextWatch for subsequent empirical studies based on this psychological framework validating the quantitative impact of the three-utility model on actual XAI adoption rates.Importance 50/100