Better Together? The Role of Explanations in Supporting Novices in Individual and Collective Deliberations about AI

Group deliberation creates shared understanding that individuals miss: While individual participants engaged more deeply with explanations and performed better on study tasks, group settings enabled participants to collectively build understanding and develop balanced arguments both for and against AI system deployment—a benefit unavailable to those working alone.

Modular explanations with four information categories work for novices: The study found that AI explanations organized into four distinct information categories effectively supported both individual and group decision-making among people with no technical background, suggesting a practical design approach for public sector AI discussions.

Explanations alone aren't enough for meaningful public deliberation: The research highlights that while explanations help novices understand AI systems, they work best when combined with group discussion settings where stakeholders can exchange perspectives and collectively weigh the merits and harms of deploying AI in public institutions.

Group settings need specially designed explanations: The paper provides specific recommendations that explanations should be redesigned to function effectively in group contexts, not just for individual users—suggesting that current XAI approaches may not fully support the collaborative deliberation needed for responsible AI deployment decisions.

Public participation in AI decisions requires accessible tools: The findings underscore that enabling non-technical stakeholders to meaningfully participate in decisions about AI systems in public institutions requires both clear explanations and structured group deliberation opportunities, not just technical documentation.