When the Machine Teaches the Human to be Human Centered

As my friend Dan Gilbert says, "design thinking is about making someone's life better." In using design thinking to make someone's life better, one needs to get to know that person. In the design thinker's toolkit, there are many techniques to uncover a person's unmet needs. One powerful way to do that is to have a great conversation.

Teaching Empathetic Conversations

I teach empathetic conversations in my design thinking class at the University of Kentucky. Getting these just right can be tricky for my students who are new to design thinking. One has to be attuned to listening for keywords in a person's response that may indicate hidden pains, desired gains, or points of frustration or hope.

As my students engage users in a conversation about a challenge we're addressing, I implore them to make the pains or desired gains users are experiencing as concrete as possible. Osterwlader et al. explain this well in their book Value Proposition Design (p. 15).

"...when a (person)?says 'waiting in line was a waste of time,' ask after how many minutes exactly it began to feel like wasted time. That way you can note 'wasting more than?x?minutes?standing in line.' When you understand how exactly (people)?measure pain severity, you can design better pain relievers in your (solution)."

Students get better at doing this with practice. But that practice usually comes by holding an empathetic conversation with people in the context of a project. That means students may make more mistakes in early interviews than in later ones.

Using ChatGPT as a Partner in Deliberate Practice

So how do I help mitigate that? Currently, I share tips from other authors, like Osterwalder et al. (e.g., what do they mean when they say "that costs too much?" How do they define too costly? What do they mean when they say "that makes me happy? How? In what way?). I also ask my student to watch a video from the IIT Institute of Design on "Getting People to Talk: An Ethnography & Interviewing Primer." I also have them participate in an empathetic conversation role-play from my book,?Design Thinking in Schools.

While these learning resources help, they are not?deliberate practice.

Ethan Mollick recently wrote about?a way to simulate deliberate practice using AI, specifically ChatGPT. And what he concluded is this:

Even in its current form, ChatGPT is shockingly close to being able to help anyone, anywhere learn via deliberate practice.

Replicating the Experiment

I replicated Mollick's experiment using ChatGPT to teach negotiating skills by asking it to teach empathetic conversation skills.

I gave ChatGPT this prompt: I want to do deliberate practice about how to conduct empathetic conversations with users for the purposes of discovering unmet needs and latent needs they have. You will be my teacher. You will simulate a detailed scenario in which I have to engage in an empathetic conversation with a user . You will fill the role of one party, I will fill the role of the other. You will ask for my response to in each step of the scenario and wait until you receive it. After getting my response, you will give me details of what the other party does and says. You will grade my response and give me detailed feedback about what to do better using the principles of user experience design and design thinking. You will give me a harder scenario if I do well, and an easier one if I fail.

Here is the result:

This was better than I expected and similar in quality to what Mollick attained with his negotiation training scenario.

What I Like, As a Teacher

ChatGPT knew, just from my request, that it behaves like a UX designer and a design thinker, that I should ask open-ended questions, reflect on my user's responses, and use that reflection as the basis for choosing my next question. ChatGPT used a common term of art in our work to "dive deeper" and stressed that I should focus on Sarah's experiences and needs.

What Needs Work

I can't fault ChatGPT for anything it did. While it didn't "grade" my response as asked, this part of the prompt may have been vague. I did ask for "detailed" feedback. I think it would be nice if the feedback were more detailed.

This meant the work is on my part to craft a better prompt.

For instance, I would like to try to craft a prompt that asks ChatGPT to:

• Encourage the learner to use the five-whys technique
• Catches the learner when they don't follow up to define ambiguous responses (like "costly" and "happy").
• Offer more in-depth feedback at the end about what insights could be gleaned from her responses

What I Want to Do Next

I want to test this approach with my students. Before I do, I want to re-engineer the prompt to see if I can enhance the feedback and see what the responses are if I do not take ChatGPT's advice of using open-ended questions, miss asking questions about needs or experiences, or don't follow up on potential pain points or gain enablers.

That these results are attainable now with the free public demonstration version of ChatGPT is amazing to me. It broadens my thinking on how AI can enhance my creativity and effectiveness as a teacher. Even at this early stage, I have new interactive ways of teaching tricky skills to my students. That's exciting to me.