How to be a design student in the age of AI

So we have this great new tool that has a big effect on the way we can design products. It sometimes stuns, it sometimes creates really disappointing work. In this post I’m trying to formulate how I think you could use the tool as a design student. Let’s start with the endgoal: At the end of design education you should be a good designer. A person who is able to structure open problems, and who has the technical skills to make high quality and effective “artefacts”. Every tool that you use during this process should serve this goal. Every tool should help you to become a better designer. So what opportunities does AI give that make you a better designer?

Apply theory

Lets start with applying theory. You might have had courses that gave you interesting theories that you might be able to apply in design problems. First of all: a conversation with a good chat agent helps you to refresh your the knowledge about the theory. Make sure that you are in charge, state how you think the theory works. Ask for questions, ask for feedback. Let it help you to start your thoughts. Treat it like an good inspiring friend, who sometimes is full of shit. Another use is to get reacquainted with knowledge you have floating around in your head: who was it again that thought design paradoxes are useful for finding frames.

Explore futures

When you have some theory, you might use the AI to explore multiple possible futures. For example: how could I apply the theory of planned behaviour to help students pick more sustainable products? What other parties could I reach with this theory, and what would it look like? Use this to find other dimensions that matter. Use the results as inspiration, not as the truth.

Coaching

Design is a multifaceted job, and when you are a design student it is even more complicated because you also have to prove your progress to a different audience. Let AI coach you. A first thing you could do is explain your approach, and ask for criticism. You can do this for steps in your project, but also for code. Make sure that you ask follow up questions. A second thing that you can do is give AI insight in the requirements of your education (like competencies or learning outcomes). You could copy and paste a learning outcome for your university, show your material and ask for feedback. What works? What doesn’t? And why.. When you let AI coach you, a big benefit will be the rubber ducking process. Rubber ducking is the famous process where you are allowed to ask your collegues any question, but only after you have had a conversation about the problem with a rubber duck. By just explaining your problem, you might just find the answer. Just formulating the questions and being aware of what you’re looking for is often more useful than the response. Like Charles Kettering said “A problem well stated is half of the solution”

For example: You created evidence that you are proficient in HTML/CSS, and want to make sure that you can prove it to your teacher. Ask what a good way could be to show your evidence. And also ask in what way you can test the quality of your work.

You applied design thinking to your project. You have created all kinds of artefacts and what not. Ask for questions about your process. Ask what could be a good way to show the process without going to much in the weeds. While debugging a product, explain in detail what the problem is. Describe what steps you want to take, ask for feedback.

Prototying

If you are developing a product in your project. You might need some throwaway prototypes to test ideas. You can give detailed input to a generative AI to generate code for a script. For example: generate a one page website that uses strategy X to influence students to make more sustainable choices. The code can be used to test whether a certain strategy holds water. Generating can also be useful when creating more test data. For example, in day planning, here’s a to-do that a client could have: “🪥Brush teeth”. Now come up with 10 extra todo’s.

Generative AI could also function as a placeholder. When you are approaching a complex technical problem, you might not have enough time to make all the stuff that is required. Or you might not have a complete picture of what is needed. “I have done several projects in which we used OpenAI’s API to simulate a backend.. For example: for an app that helps people with a visual disability we instructed open AI to come up with a fitness program that fits within a users question formatted in JSON. The JSON part made it readable for an app. For testing purposes open AI’s answer was good enough. Enough to see whether there is enough here to justify investing more time.

Cut corners

Every now and then cut corners: Let say there is this one screen in your app that just shows debug data. You can generate part of that code. But if you plan to ship it, make sure that you understand what it does. Don’t claim ownership, and review it for maintainability.

Conclusion

AI has definitely already changed how we work as designers and developers. Taking the lazy approach will not help you in the long run. It might get you through school. But in the end: what will be skills that an employer hires you for? What do you bring to the table? Make sure that you have a good answer for these questions.