Productive Discussion Moves in Groups Addressing Controversial Issues

Emotions and personal experiences strengthen ethical discussions: The research found that when students incorporated emotional and experiential arguments into their discussions about AI ethics dilemmas, it was associated with higher-quality conversations—challenging the common assumption that good reasoning requires removing emotion from the equation.

Acknowledging complexity is a sign of productive dialogue: Explicit acknowledgment of ambiguity and the difficulty of ethical questions emerged as a strong positive predictor of discussion quality, suggesting that recognizing the genuine complexity of AI ethics issues leads to more substantive conversations than seeking definitive answers.

The sequence of discussion matters as much as individual contributions: Using Ordered Network Analysis, the study found that productive discussions followed specific patterns where emotional expressions were connected to evidence-based reasoning, indicating that how ideas are sequenced and linked together—not just what is said—shapes discussion quality.

Building on others' ideas alone isn't enough: Surprisingly, the study found that simply building on previous ideas was negatively associated with discussion quality, suggesting that productive ethical discussions require more than agreement or incremental agreement—they need substantive engagement, reasoning, and sometimes constructive disagreement.

Current understanding of productive discussion is STEM-biased: This research addresses a gap by showing that productive discussion moves in ethical and value-laden contexts (like AI ethics) differ from those identified in science and math discussions, highlighting the need for context-specific approaches to facilitating responsible AI conversations.