AI’s Impact on Design Ideas: A Hands-On Exploration Workshop at UX Camp 2024
Jacynthe Roberge, Associate Professor, Design, Laval University, Québec, Canada
Isabelle Sperano, Associate Professor, Design, MacEwan University, Edmonton, Canada
Vik Chu, Design Student, MacEwan University, Edmonton, Canada
At the end of 2022, the striking arrival of ChatGPT caught many of us off guard. Its launch, coupled with its seemingly efficient performance, was an unexpected wake-up call to the potential impacts AI could soon have on multiple professions and activities. Two years on, several applications of generative AI have found their way into our daily professional routines, shaping various sectors with a growing influence and tackling tasks we hadn’t anticipated it would address so soon, to the point where recent developments in AI suggest it can revolutionize creativity, especially regarding idea generation (De Cremer et al, 2023).
As design professors — Jacynthe Roberge is the director of the Interaction Design Masters degree at Laval University, while Isabelle coordinates the Digital Experience Design programs at MacEwan University — this question intrigues us. We decided to dig into this topic during a workshop held at UX Camp Edmonton 2024, an established yearly event organized by UX Edmonton. The title of our workshop was “The Impact of AI on Design Ideation: A Practical Exploration” and aimed to explore the potential impacts of AI on participants’ idea generation and creativity during the research phase of an innovation-driven UX design project. To do so, participants were asked to collaborate in a short, open-ended design challenge targeting the first phases of the UX design process and, then, to engage in critical reflection and discussions about the benefits and pitfalls of integrating AI into the UX ideation process based on their experience in the design challenge they were given during the workshop.
Workshop Structure and Progress
To achieve this goal, the workshop was structured as follows:
- Introduction and Instruction
- Design Brief Unveiling
- Team Creation
- Design Challenge Execution
- Results Presentation, Experience Comparison and Discussion.
1. Introduction and Instruction
To ensure consensus on the overall procedure, we first introduced the workshops, reviewed the UX design process, and established a shared design model comprising four key stages characterized by divergent and convergent thinking activities: immersion, framing, conceptualization, and design. While all participants were encouraged to adhere to this process, there was no obligation for exhaustive completion. The aim was not to foster a competitive or efficiency-driven mindset, but rather to encourage observation and reflection on the nature and quality of the creative activities inherent in our design approaches.
2. Design Brief Unveiling
We then presented the design challenge participants would have to work on. In alignment with the eleventh of the UN Sustainable Development Goals, participants were tasked with generating solution ideas and developing UX design concepts that make the city of Edmonton inclusive, safe, resilient, and sustainable.
Their final design ideas needed to be user-centered, creative, original, and effective. To help participants, we provided a list of generative AI applications for inspiration, without requiring their use.
3. Team Creation
To observe the effects of AI on participants’ creativity and idea generation, we divided the group into three distinct types of teams:
- AI-free: participants were restricted from using AI (1 team).
- AI-supported: humans and AI collaborated (2 teams).
- AI-driven: AI took the lead in guiding the human designers (2 teams).
Fourteen professionals, both designers and non-designers, participated in the workshop. Four design student volunteers helped us during the workshop. They all participated in the design challenge as an active member of a team. They were tasked with keeping the teams actively working on a common goal and reporting their team’s experiences at the end of the session. Only the AI-free team did not have a student member.
At the beginning of the session, we did a quick survey about the participants’ perceptions regarding the integration of AI into the field of UX design and its impact on designers’ creativity. Most of the participants’ responses were very positive, and the group seemed particularly enthusiastic about AI overall.
4. Design Challenge Execution
The teams had approximately 90 minutes to complete their design challenge. The design model presented at the beginning of the workshop served as a collective guide, encouraging participants to immerse themselves in the project, frame their objectives, conceptualize solutions, and develop prototypes. During the design challenge, participants were invited to use Miro to record their ideas and keep track of their design process.
Just before letting them begin their design activity, we asked participants to rate their creative confidence (Will they generate few or many ideas? Will these ideas be of good or bad quality?) and their expected experience (Will the experience be rather easy or difficult? Will it be rather pleasant or unpleasant?). At the halfway point, as well as at the end of the design challenge, we repeated the task and invited participants to reflect on these same criteria.
5. Results Presentation, Experience Comparison and Discussion
By the end of the challenge, each team presented their key ideas and concepts, shared their design approach and compared their different experiences. They reflected on their creative journey, discussing what worked, what didn’t, and why. Lastly, we all engaged in a discussion regarding the potential benefits and pitfalls of integrating AI into the UX Ideation process, drawing insights from their experience during the workshop’s design challenge.
What We Thought Would Happen
Before starting the workshop, we had some assumptions about how things might unfold and the experience each type of team would have.
AI-Free Team
Given their lack of reliance on AI assistance, the challenge may prove to be more demanding for this team. Without access to generative AI tools, their proposals will likely draw from personal experiences, knowledge, and values. Given that the time allocated to the exercise is short and that the participants might need some time to exchange and share their ideas, this team might therefore run out of time and not succeed in going through the whole design process. They might work at a slower pace and generate fewer ideas than their AI-assisted and AI-driven counterparts. However, the quality of their ideas should not be compromised by the absence of AI.
AI-Supported Teams
This type of team seemed likely to have the most advantageous position among the three as they have the freedom to utilize AI as they think necessary. AI would likely function as a highly efficient team member, capable of rapid idea generation and information retrieval. In addition to being efficient, these teams would likely have more time to focus on developing high-quality ideas. Both the quality and quantity of ideas were anticipated to be integral aspects of their projects.
AI-Driven Teams
By allowing AI to lead the project, we assumed that these teams would easily generate a vast number of ideas and make rapid progress in the design challenge. However, there were concerns regarding AI’s ability to discern high-quality creative ideas, doubts lingered regarding the relevance of the solution ideas that would be produced by AI-driven teams.
What Actually Happened
After discussing and debriefing with the teams, we drew some insights and observations regarding the progression of the design challenge as experienced by our three types of teams (AI-Free, AI-Supported, and AI-Driven). Below, for each team type, we summarize our observations regarding their design process (including their project status, solution ideas, and concept development progress), perceived experience, and views on the creativity of their project.
AI-Free Team
- Project status, solution ideas, and concept
The AI-Free team was much quicker than we had anticipated. Despite being slightly slower and less efficient than the AI-driven team, they nearly matched the pace of the AI-supported teams. They were at the Conceptualization stage when the challenge ended. The AI-Free team successfully explored two concepts they found interesting and had time to develop one further into a digital product concept with multiple feature ideas. - Perceived impact on experience
The team members shared that they thoroughly enjoyed the workshop experience, finding it to be enjoyable and collaborative. Their teamwork was evident not only in their collective approach to problem-solving but also in their synchronized responses to the short mid- and end-of-challenge exercises that we asked them to do (evaluation of creativity and lived experience). Although they emphasized having missed the support of AI, particularly for secondary research, overall, they appreciated the learning and bonding opportunities which stemmed from the nature of their team. - Perceived impact on creativity
In the first few minutes of the exercise, the team said they consciously embraced creativity by allowing themselves the freedom to entertain “silly ideas.” This openness to unconventional thinking fostered a process characterized by both divergent and convergent thought patterns. Their collaborative efforts were instrumental in nurturing creativity, with team members acknowledging and valuing each other’s contributions throughout the workshop.
AI-Supported Teams
- Project status, solution ideas, and concept
One team had progressed to the Conceptualization phase, while the other had reached the Design phase. They seem to have easily adjusted to the time allocated for the exercise. The transition through these stages occurred smoothly, with no notable issues or setbacks encountered. The first team went to the stage of exploring different concepts, while the second team developed a single concept with a slightly deeper level of detail. - Perceived impact on experience
Several common challenges were encountered by both teams. For example, they experienced moments where they forgot to use AI. They struggled with knowing when and how to use AI tools effectively. For example, although one of the teams was having difficulty refining its ideas, it did not have the reflex to consult the AI to do so. This suggests that the added value of AI lies, among other things, in the planning of the design activity and the prior knowledge of the tool’s various benefits for design. Initially, they relied on AI for divergent thinking (exploration, idea generation), but had to consistently remind themselves as a team to utilize AI throughout the process. Most teams found AI most beneficial for tackling uncertain or unfamiliar topics, treating it as an expert advisor in those areas. - Perceived impact on creativity
Both teams utilized ChatGPT to generate solution kernels (ideas). Some participants mentioned that using Chat GPT helped them generate key ideas they wouldn’t have thought of otherwise and allowed them to explore these ideas as a starting point for creating concepts they wouldn’t have discovered on their own. According to one of the teams, generative AI-enabled them to reach a level of creativity they wouldn’t have achieved otherwise within this time frame, as it helped broaden their general knowledge and solution repertoire. One team used ChatGPT to assess the quality of their ideas. While acknowledging that their solution wasn’t particularly original, they expressed confidence in its effectiveness.
AI-Driven Teams
- Project status, solution ideas, and concept
The projects progressed really quickly. AI-Driven teams went far in the design process. Although both teams managed to present very elaborate concepts, they arrived at very different results from each other. The first team let the AI take the lead on everything — from the project’s direction to the decisions that had to be made. While doing so, they didn’t make the AI stick to a conventional design approach. Therefore, they ended up with a very broad solution that looks like a manifesto for a perfect society. The other team let AI guide them while directing it in the direction of the conventional design approach. They did a lot of convergent and divergent thinking, and by the end, they had three detailed UX design concepts in hand — something that would usually take weeks of work. They managed to generate interfaces, a draft budget, a hiring plan for the necessary team, and more. In addition to having developed a complete concept, one of the teams went as far as creating the presentation of their concept as if they were giving a TED Talk. - Perceived impact on experience
While many participants were pleasantly surprised by the amount of work accomplished in such a short period, none of them felt a sense of ownership over their ideas. One of the teams had not yet seen the result of their final solution before presenting it to the rest of the group. When you let AI take control, it moves forward, but your design experience and control over the project are certainly affected. Also, they didn’t know if the ideas were valid or not. - Perceived Impact on Creativity
The speed of the process prevented participants from properly assessing the quality of ideas generated by AI. While these projects undoubtedly generated a significant number of ideas of all types (project, problems, opportunities, solutions, functionalities, interfaces, budget, etc.), there was a lack of means to assess their quality because they were not knowledge experts and couldn’t fully trust the AI’s sources and summaries. The resulting solutions were either excessively broad, superficial, or overly precise, often lacking a clear rationale for their selection given the particular setting in which the design challenge took place.”
As mentioned before, three times during the design challenge, participants were asked to assess their creative confidence and their anticipated experience quality. These brief surveys helped us track the evolution of their perceptions on these two aspects and provided insights into their progress throughout the challenge.
Insights and Lessons Learned
Importance of Clear AI Guidelines for the Integration in a UX Design Process
Reflecting on our workshop experiences, we recognized the potential benefits of the integration of artificial intelligence (AI) within the UX design process. However, without explicit planning and prior consideration of AI’s role in the UX design process of a project, its potential may be overlooked or underutilized. Instances where AI tools were forgotten or could have been used more effectively highlighted the importance of establishing clear guidelines and frameworks for AI integration in UX design workflows to help designers understand how, when and why to incorporate AI into their design methodologies.
Creative Use of AI in UX Design
Throughout the workshop, some teams touched on inventive applications of artificial intelligence (AI) within UX design. For instance, one team crafted a TED talk-style presentation to convey their solution, showcasing AI’s potential not just as a problem-solving tool but also as a creative medium for communicating design concepts. This experience underscored the importance of pushing the boundaries of traditional design methodologies and embracing AI’s capacity to inspire fresh perspectives and innovative approaches in UX design.
Broadening the Designer’s Repertoire
Designer’s repertoire refers to the set of known responses, within the designer’s cognitive framework, available to a designer when faced with a design problem (Schön, 1983). It allows designers not to reinvent the wheel each time they face a new problem, but rather to rely on proven solutions and strategies. By using generative AI to suggest problems and propose solutions, participants were exposed to alternative perspectives that they could not previously consider because they were not part of their preexisting repertoire.. This quickly pushed beyond their existing knowledge and encouraged them to break free from conventional thinking patterns, considering solutions outside their conventional solution repertoires. Although a more in-depth literature search and precedent analysis could have potentially produced similar results, this was not possible within the scope of this design challenge given the limited time available for the exercise.
Enhancing Idea Generation with AI
This workshop highlighted the capacity of AI to generate a vast quantity of ideas, significantly enriching ideation generation. However, it’s crucial to recognize that while AI can assist in idea generation, the responsibility for validating, evaluating, selecting, and ultimately assuming ownership of the ideas remains with the human designer. It’s important for humans to always evaluate and validate AI-generated content so the limitations and flaws of AI (e.g. hallucination) must be taken into account. While AI can augment ideation with its computational capabilities, human oversight and validation are indispensable for ensuring that the ideas align with the project needs and goals, and for providing a holistic perspective that AI alone may lack. It’s imperative for designers to retain control over the ideation process, ensuring that human judgment and creativity guide the selection and refinement of ideas. By maintaining this balance in the human-AI interaction, designers should be able to harness the potential of AI while preserving their responsibility in the creative process.
Conclusion
We think this workshop provided the group with a hands-on exploration of AI for creativity and a possibility to discuss this experience. Human-AI interaction is still relatively new to all of us, and there is much to explore, discover, and learn as we navigate the evolving landscape where human creativity and AI intersect.
While AI seems to have shown some ability to assist in idea generation, it also presents challenges concerning creative control, maintaining originality and verifying the accuracy and truth of information. As AI technology continues to advance, its integration into UX design will necessitate careful consideration and a delicate balance between leveraging its capabilities, and preserving human control and a sense of ownership.
Acknowledgments
We would like to extend our heartfelt thanks to all the participants who made this workshop a success.
- Alyana Angus
- Tanya Camp
- Korapin Chaotakoongite
- Dallas Dyson
- Jacqui Frechette
- Juhi Gupta
- Kartik Gupta
- Mark Hill
- Tony Joseph
- Igor Keleberda
- Linnea Lapp
- Jessica Malik
- Nathan Smith
- Angela Wong
We also warmly thank the students who generously agreed to lend us a hand in setting up this workshop.
- Nawaal Basha
- Vik Chu
- Danielle McDow-York
- Soyinka Seguin
References
DE CREMER, D., BIANZINO, N. M. et FALK, B. (2023). « How Generative AI Could Disrupt Creative Work », Harvard Business Review, https://hbr.org/2023/04/how-generative-ai-could-disrupt-creative-work.
SCHÖN, D. (1983). The Reflective Practitioner : How professionals think in action, Basic Books, 374 p.
Contact
For more information, please contact Jacynthe Roberge or Isabelle Sperano.
