Using AI to create the practice opportunities online students need

The most valuable thing artificial intelligence gave my students wasn’t answers. It was practice.

For years, I faced a challenge familiar to many instructors who teach asynchronously online. My students were learning the concepts, reflecting deeply and developing new perspectives, but they weren’t getting enough opportunities to practise the interpersonal skills they would need in their careers.

As a psychologist by training who teaches leadership and conflict resolution in sustainability-focused programmes, much of my work centres on helping students navigate complex environmental challenges. My students must learn to communicate across differences, lead in uncertain situations and engage productively with people whose values and perspectives differ from their own. These are the same skills they will rely on in workplaces, communities and public decision-making.

The need for practice

These are not skills that can be mastered by reading about them, though. They require practice.

This is best explained by an example. In one of my courses, students work through exercises that help them examine real conversations they have experienced in highly polarised environments. Often, these conversations involve climate change, biodiversity or environmental policy. Surprisingly, the most emotionally charged examples are frequently not workplace interactions but conversations with family members, room-mates or close friends – in relationships that are consequential for them.

• How an AI-integrated anonymous chatroom boosted AI literacy and inclusion
• What future-proofing a course really means
• How to be present in teaching, even when you’re not physically there

Many students begin the course believing that success means winning an argument. They want to persuade the other person, change their mind, prove that they are right. Through reflection and guided activities, they gradually discover that meaningful communication often looks very different. Success may mean preserving a relationship, keeping a conversation going or creating conditions for future understanding rather than immediate agreement.

The learning can be transformative. Students regularly tell me that these experiences change the way they approach difficult conversations. But the problem is that knowing what to do is not the same as being able to do it.

In a traditional classroom, students might role-play these conversations with classmates. They could test communication strategies, receive feedback and refine their approach. In an asynchronous online course, however, creating those opportunities is far more difficult.

I tried offering voluntary role-playing sessions. Students could meet with one another or with me to practise these conversations, but very few participated. I understand why; many online learners are balancing careers, families and competing responsibilities. Scheduling additional meetings can be difficult. Yet the need for practice remained. 

An AI experiment that paid off

That challenge led me to experiment with generative AI. The goal was never to replace human interaction, nor was it to showcase the latest technology. My goal was simply to create an environment where students could safely experiment with difficult conversations.

I developed two AI-powered tools. The first served as a role-play partner that held different perspectives and, importantly, was programmed to resist easy persuasion. Some scenarios were moderately challenging for students. Others were intentionally difficult. The second tool served as a coach. Drawing on the communication and emotional intelligence skills students had been studying throughout the course, the coaching tool provided feedback on how effectively students navigated the interaction.

Building these tools took months of developing, testing and refining. Every instruction mattered. I repeatedly adjusted prompts, corrected unintended behaviours and ensured the feedback aligned with course objectives. Before students encountered the system, I enlisted colleagues to help test it as well.

The result surprised me. Students did not simply interact with the AI, they meaningfully engaged with it. Many reported that the experience felt valuable precisely because the simulated conversation partner pushed back. Unlike classmates who sometimes hesitate to challenge one another, the AI consistently maintained its position. Students had to work harder, listen more carefully and apply the communication strategies they had spent weeks learning. And, importantly, they practised.

What surprised students most was the quality of the coaching. Emotional intelligence is often considered uniquely human, but I had programmed the AI tool, so it could provide feedback about how they communicated, listened and responded to conflict. 

The future of AI in asynchronous – and synchronous – education 

The experience raised important questions about the future of AI in education. What if AI's greatest educational value lies not in generating answers but in generating experiences?

For disciplines that depend on communication, leadership, negotiation, mediation and other interpersonal skills, AI can create opportunities for practice that would otherwise be difficult to provide, particularly in asynchronous environments.

Looking ahead, I can imagine even richer possibilities. Students might interact with AI-powered stakeholders in environmental conflicts, practise mediation skills or engage with simulations that adapt dynamically to their decisions. Visual avatars could add another layer of realism. Scenarios could become increasingly personalised.

We should recognise that many students approach AI with mixed feelings. In sustainability programmes especially, students are often keenly aware of concerns surrounding energy consumption and water usage. Others worry about how AI may affect future employment. As educators, we cannot ignore these concerns. If we ask students to engage with AI, we should also create space for critical conversations about its limitations, trade-offs and impacts. 

The most valuable thing AI gave my students was not information. It was something far more difficult to provide in an online course: the chance to practise, make mistakes, receive feedback and try again before the stakes become real. By the time students encounter these conversations in their workplaces, communities or families, they have navigated them in a low-risk environment where learning, not winning, is the goal.

Kathy Miller Perkins is professor of practice in the  Center for Leadership in Global Sustainability at the College of Natural Resources and Environment at Virginia Tech. She is owner and CEO of a consulting firm and the author of Leadership and Purpose: How to Create a Sustainable Culture (Routledge, 2019).

If you would like advice and insight from academics and university staff delivered direct to your inbox each week, sign up for the Campus newsletter.