Designing Better Discussions in an AI World

Discussion boards have always been challenging. Long before generative AI arrived, many instructors struggled with discussions that felt repetitive, superficial, or disconnected from meaningful learning.

Students often learned the formula quickly:

• Read the prompt.

• Write a response.

• Reply to two classmates.

• Move on.

AI introduces new complications. When students can instantly generate polished responses, discussions can quickly become collections of well-written but shallow contributions. Everyone appears to be participating. Yet very little meaningful thinking may be occurring.

When I hear faculty express concerns about AI and discussions, the instinct is often to make prompts harder. More complex questions. Longer responses. Additional requirements. But I'm not convinced difficulty is the solution.

I think the solution is better prompts. More specifically, prompts that focus less on producing information and more on exploring ideas. Because the best discussions have never been about generating answers. They've been about making sense of ideas together.

That's an important distinction.

AI is increasingly capable of producing answers. It is much less capable of engaging in the kinds of social learning experiences that occur when people wrestle with uncertainty, challenge assumptions, and refine their thinking through interaction with others.

When discussions focus primarily on summarizing information, AI can often do much of the work.

But when discussions ask students to:

• critique ideas

• evaluate examples

• connect concepts to personal experience

• compare perspectives

• identify assumptions

• explain reasoning

• wrestle with uncertainty

something different happens. The discussion becomes less about information and more about judgment. And judgment is much harder to outsource.

This connects directly to an idea I've explored throughout this series: making student thinking visible. Discussions can be powerful because they provide opportunities for students to externalize their thinking. Students explain their reasoning. Defend their positions. Respond to alternative viewpoints. Clarify misunderstandings. Revise their conclusions.

All of these activities make thinking visible. But discussions can provide something else as well. They can help students examine their thinking in relation to other people's thinking.

Educational researchers sometimes describe this process as social metacognition. Reflection helps students think about their own thinking. Discussion helps students think about their thinking alongside others. When students encounter perspectives that differ from their own, they are often forced to examine assumptions they didn't realize they were making.

When they explain an idea to someone else, gaps in understanding frequently become more apparent. When they hear an alternative interpretation, they may discover new ways of approaching a problem. These moments matter because learning is not always an individual activity.

Many of the decisions people make in professional settings emerge through conversation, collaboration, questioning, and collective problem-solving. Discussion creates opportunities to practice those skills. And those opportunities may become even more valuable as AI becomes more integrated into our work and learning environments. Because AI can provide answers. But it cannot fully replicate the experience of negotiating meaning with another human being. It cannot replace the experience of discovering that someone else interpreted the same information differently. It cannot fully reproduce the intellectual growth that occurs when our assumptions are challenged by another perspective.

That doesn't mean discussions should ignore AI. In fact, AI may create opportunities for what I think of as synthetic social metacognition. Students might ask AI to challenge their assumptions, argue an alternative perspective, or identify weaknesses in their reasoning before engaging with classmates. The goal isn't for AI to replace human discussion. It's to create additional opportunities for students to examine their own thinking. Used this way, AI becomes less of an answer generator and more of a catalyst for reflection.

The goal is to ensure students remain active participants in the thinking. Because ultimately, the value of discussion has never been the discussion post itself. The value lies in the intellectual interaction the discussion creates. The exchange of ideas. The examination of assumptions. The refinement of understanding. The development of judgment.

Those outcomes remain important regardless of how capable AI becomes. And when discussions are designed well, they become another powerful way to make student thinking visible.

But visible thinking alone isn't enough. Students also need clarity about expectations. They need to understand what role AI should play in their learning and what responsible use looks like within a particular course.

That's where transparency enters the conversation.

Series 1: AI Is Exposing Existing Problems ✓ Completed

Series 2: What We Do About It

Current Post (6 of 8): Designing Better Discussions in an AI World

Next Up: From A! Rules to Learning Goals