On the Influence of Cognitive Styles on Users' Understanding of Explanations

One-size-fits-all explanations don't work for everyone: The research demonstrates that people with different cognitive styles—specifically rational versus intuitive thinkers—understand AI explanations differently. This suggests that effective AI transparency requires tailoring explanations to how individual users naturally process information, rather than assuming a single explanation format will work for all users.

Cognitive psychology should inform AI design: The study highlights a critical gap in current explainable AI (XAI) approaches, which have focused primarily on technical solutions while overlooking how human cognitive differences affect comprehension. Incorporating insights from cognitive psychology into XAI design could significantly improve users' actual understanding of AI decisions.

Both objective and subjective understanding matter: The research measured not just whether users correctly understood explanations (objective understanding) but also how confident they felt in their understanding (subjective understanding). These can differ based on cognitive style, meaning users might feel confident about an explanation they don't actually understand—an important consideration for responsible AI deployment.

User-centered design is essential for trustworthy AI: By taking a human-centered perspective rather than a technology-focused one, this work emphasizes that responsible AI requires understanding the people who interact with these systems. Explanations that fail to account for cognitive differences may undermine trust and lead to poor decision-making by users relying on AI systems.