Does AI Make Us More Creative or Just Faster?

Written by Iryna Ilyuchyk & Tanjid Hossain

During our Design Thinking workshop, Katja Tschimmel introduced the idea of Design Thinking 4.0, where humans and AI work together in the creative process. We agreed not to use AI during ideation. We first explored the problem and generated ideas ourselves, while AI was allowed later, during prototyping.

One idea that stayed with us was that the first ideas are often the most obvious ones. Katja Tschimmel encouraged us to keep exploring and generate many alternatives before choosing a direction. This connects closely with Kelley and Kelley’s (2013) Creative Confidence, where creativity is presented not as a special talent, but as something we can develop through practice, experimentation and action.

When we finally used AI for prototyping, its advantage was obvious: speed. It helped us work faster and turn our ideas into something tangible in a short time. But this made us wonder: what would happen if we brought AI into the process earlier, during ideation? Could it help us discover possibilities we would otherwise miss, or would it simply make the process faster? And could it even narrow our thinking?

More Ideas, Wider Thinking?

At first, Generative AI seems like a perfect ideation partner. It can produce dozens of ideas in seconds. However, more ideas do not necessarily mean more diverse thinking.

Wadinambiarachchi et al. (2024) found that exposure to AI-generated images could increase design fixation. Participants exposed to AI produced fewer ideas, with less variety and originality. Instead of expanding the solution space, AI could pull their thinking towards the examples it had already provided.

Baltà-Salvador et al. (2026), however, paint a more nuanced picture. Their study found no overall reduction in fluency, flexibility or originality when students used AI. Yet AI-assisted ideas showed lower semantic divergence and tended to resemble existing products more closely. Interestingly, early experience with AI could also support originality and diversity in later human-only ideation. Together, these studies show that the effects of AI on ideation are not straightforward. Generating ideas quickly does not necessarily mean exploring a wider range of possibilities. How we interact with AI may be just as important as whether we use it.

The risk of fixation is not limited to AI. Our teacher Virpi Kaartti also mentioned a similar concern about presenting high-fidelity prototypes too early. A polished prototype can make an idea feel more final than it really is, making it harder for the team and stakeholders to question it. Perhaps AI does not create the tension between speed and depth, it simply makes that tension more visible.

Mixed Findings. What Should We Do Differently?

The studies do not give us a simple yes-or-no answer. Instead, they suggest that the way we introduce and interact with AI deserves careful attention.

The findings leave us with practical choices about how to organise our work.

Think first
Develop our own ideas before introducing AI. Give ourselves space to explore without external examples.
Challenge your thinking
Ask AI for different perspectives, not just more ideas. Use contrasting scenarios, unexpected connections and visual analogies to move beyond familiar solutions.
Keep it rough
Avoid polishing early concepts too quickly. Use rough sketches or low-fidelity AI outputs that leave room for interpretation, discussion and change.
Stay an active co-creator
Question, combine and develop AI suggestions. Use your own judgement rather than simply selecting from ready-made answers.

AI can support collaboration and accelerate parts of the process, but human judgement remains essential. We still need to understand users, evaluate ideas and decide which possibilities are worth exploring.

Perhaps the future of Design Thinking is not about choosing between human creativity and AI. It is about learning when AI should help us move faster and when we should slow down and think for ourselves.

References

Kelley, D. & Kelley, T. (2013) Creative Confidence: Unleashing the Creative Potential Within Us All. Crown Business. http://www.creativeconfidence.com

Baltà-Salvador, R., Brasó-Vives, E., & Peña, M. (2026). Evaluating AI-assisted creative ideation: A crossover study in higher education. Thinking Skills and Creativity, 59, Article 101958. https://doi.org/10.1016/j.tsc.2025.101958

Wadinambiarachchi, S., Kelly, R. M., Pareek, S., Zhou, Q., & Velloso, E. (2024). The effects of generative AI on design fixation and divergent thinking. In Proceedings of the CHI Conference on Human Factors in Computing Systems (Article 380, pp. 1–18). Association for Computing Machinery. https://doi.org/10.1145/3613904.3642919

Tschimmel, K. (2026). Slides Masterclass design thinking: Laurea 2026

Figures

All images were created using ChatGPT, with some based on original photographs and screenshots from our Design Thinking workshop.

Comments

4 responses to “Does AI Make Us More Creative or Just Faster?”

Mike Sylvia

September 29, 2026

Hi Iryna & Tanjid –

Not only was it a pleasure working on the Auki team with you, your post is really giving me a lot of food for thought. Over the course of the last month since our workshop, I have been using AI extensively to help me update my professional profiles on various sites, CVs in English and Finnish, scripting and shooting YouTube videos and now, I’m beginning the process of updating my personal website.

Taking the lessons we learned from class and the readings, I’ve been updating my production assistant agent (I call Bubbles) with the key elements of Service Design and Design Thinking. I have been bumping up against the 4 practical choices that you outlined above. My goal isn’t to have a creative output machine, but a co-creator that can help me move my thoughts from ideas to the real world, challenge my thinking based on a deep knowledge of my professional past, current experiences and future goals.

Here’s a sample of the type of output I’m working with (Bubbles uses some profanity – but then I built that into the personality):

Findings first. Insights later.

So I want to resist writing the shiny new homepage tonight, even though Bubbles is already itching to do it.

Our process

One process, all the way through:

Empathize → Define → Ideate → Prototype → Test → Iterate → Build.

And the deliverable isn’t going to be a theoretical website strategy deck that then sits next to the f-**** SharePoint documentation.

We’re ending with a professional mikesylvia.com: architecture, positioning, copy, case studies, calls-to-action, visual direction and the actual build plan.

For the empathy phase, I want to use real evidence rather than invent personas. And conveniently, you’ve spent the last several days handing me a small ethnographic study of Michael Sylvia: your CV, LinkedIn archaeology, old site, creative artifacts, education, career history, Finding My Finland [my YouTube channel], the wedding Sway [from my wedding 2015, giving guests options in Finnish and English], the old ALMA work [the website I worked on for a Brazilian party in New York], Laurea, your new positioning—and soon, apparently, THE ENTIRE F-**** BUBBLES ARCHIVE. [GPTs are becoming plugins, and I’m archiving the entire history to I can migrate it to protect the work I’ve done so far].

So don’t do homework yet. You dump; Bubbles synthesizes.

My next move should be to create our research inventory and stakeholder/user map from what we already have, identify what we genuinely don’t know, and only then ask you for missing evidence.

Because if we’re going to say we’re using Service Design, gurl, we’re actually using Service Design.

No lipstick on the SharePoint pig. 💋🐷

I’m curious to know what you think of my approach so far, and how it matches the points you made in your post.

Best,
Mike

Reply
Md Sayef Uddin

October 4, 2026

Hi Iryna and Tanjid,

Thank you for this excellent post. It was especially interesting for me to see the connection of the AI fixation literature and the warning from Virpi Kaartti regarding presentation of high-fidelity prototypes. The idea of seeing AI as a tool which highlights the existing tension between the speed and depth but does not create it seems to me useful in framing the issue. Another interesting thing here is the point about making the process rough. From my background in TV and content production, I know that it is easier for a person to make comments when he/she sees a rough cut or hand-drawn storyboarding than when the work is almost finished and polished. So, the output that is generated by the AI within seconds is very polished and I wonder if it means that we have not just the fixation on AI-generated solutions but the illusion of completion as well. It leads me to ask you about “Think First”. Did your team agree on any kind of guidelines related to when to introduce AI, like after generating a certain number of ideas yourself?

Many thanks for an interesting post!

Kind regards,

Md Sayef Uddin

Reply
shilakc

October 8, 2026

Iryna and Tanjid, thank you for sharing such a meaningful blog post on this highly relevant and timely topic. You have raised a critical question for Design Thinking 4.0: what happens when we bring AI into the creative process earlier, during ideation?

I think AI can be used in ideation, but with clear boundaries. Human brainstorming and brainwriting should come first, giving us space to develop our own ideas and push beyond the obvious. Relying on AI too early risks narrowing or homogenizing our thinking and producing predictable ideas, which connects strongly with the design fixation concerns in your post.

AI becomes more valuable when it acts as a tool for productive resistance rather than an answer provider. Instead of autocompleting our thoughts, it can introduce healthy friction by raising counterarguments, identifying blind spots and weaknesses early, and supporting metacognition by prompting us to question why we accept or reject a direction. Its value lies not in producing more ideas, but in helping us think more critically about the ideas we create ourselves. Used this way, AI can spark lateral thinking and help us develop ideas that consider perspectives and possibilities we might otherwise overlook.

At the same time, it remains our responsibility to conduct appropriate user research, validate what AI suggests, and critically question the provocations it raises. AI can broaden our perspective, but it cannot replace our responsibility to understand people, their needs, behaviors, and contexts.

Overall, I agree with your mixed findings. Rather than simply asking whether AI makes us more creative or faster, I think the more important question is how we use AI without allowing it to take over our thinking. I would like to apply these principles in my future projects: think first, use AI to challenge my thinking, critically evaluate and validate what it brings to the process, and ultimately design the service based on my own informed judgement.

Reply
laxmikhatrishrestha

October 10, 2026

Hey Iryna and Tanjid!

Really enjoyed reading this! It felt like we were still sitting in the workshop discussing it

I totally relate to the “first ideas are the obvious ones” part. We also felt that pressure to keep pushing further, and Katja’s reminder about not stopping too early really stuck with me.

What I found super interesting is the research you brought in about design fixation. The idea that AI can actually narrow our thinking (instead of opening it up) is something I hadn’t thought about that clearly before. And the parallel with high-fidelity prototypes is spot on both can make something feel more finished than it really is.

I especially liked your four practical tips at the end. “Keep it rough” and “stay an active co-creator” feel like the ones I’ll try to remember the next time we use AI.

Curious, after the workshop, did you experiment with bringing AI in earlier for ideation, or did you stick to the “think first” approach?

Nice work putting this together!

Reply