Silicon clay: how AI is reshaping UX design

What do the last five years of academic research tell us about how design is changing?

It would be something of an understatement to say AI has impacted the world of UX design.

But how, exactly, has it affected UX and its practitioners?

The answer isn’t straightforward, as every new AI development is accompanied by social media hype, hot takes and flexing, as designers try to prove to the world — and possibly themselves — that they know what the latest bleeding-edge model means for design.

That’s a lot of noise. How can we cut through the static to see the real picture?

Enter academia.

Over the past five years, numerous academic journal papers have been published about AI and UX design. While the wheels of academia turn slowly, we get a significant trade-off for its glacial pace: objectivity and methodical rigour.

In other words: a healthy dose of perspective.

The goal of this article is to summarise academic findings relating to four questions:

1. Where is AI being used in the UX design process?
2. What are the advantages and drawbacks of using AI in UX design?
3. How do UX design practitioners feel about using AI?
4. What are the takeaways for future use of AI in UX design?

The key sources for this article are two systematic reviews from 2025: one published in Advances of Human-Computer Interaction, and another available on the open-access scholarly platform ArXiv. In total, this article summarises findings from 17 academic sources which are referenced at the end. Typically, these studies base their findings on surveys, interviews, focus groups and other research methods to collect insights from UX and HCI practitioners.

So forget everything you’ve read from LinkedIn hype merchants, ‘thought leaders’ or doom-mongers.

Let’s summarise what the science says.

1. Where AI is used in the design process

The highest usage of AI in UX design is in the testing phase, suggests one of our 2025 systematic reviews. According to this paper, 58% of studied AI usage in UX is in either the testing or discovery stage. This maybe shouldn’t be surprising, considering generative AI for visual ideation and UI prototyping has lagged behind text generation.

Here’s what AI’s commonly being used for in each design stage:

Discovery

• Identifying design problems
• Understanding user needs and behaviours
• Creating user personas

Ideation

• Co-creating solution concepts
• Exploring design alternatives
• Predicting product values

Prototyping

• Generating UI designs
• Converting sketches to prototypes
• Checking for GUI guideline violations

Testing

• Predicting user experience
• Identifying usability issues (e.g. heuristic evaluations)
• Planning and analysing user testing

In terms of what AI tools UX practitioners are using, here’s a breakdown which is both informative but already slightly out of date:

Summary

UX practitioners are clearly making use of AI across the design process, with ChatGPT being the most popular tool.

However, while ideation and prototyping with AI gains a lot of attention, it’s the testing phase where the most studies have examined utilising AI — and it’s also significantly studied in the discovery phase. This disparity may even out as AI continues to improve at generating UI designs.

2. Advantages and drawbacks of using AI in UX design

I’ve grouped these insights from academic papers into broad themes. You can see that for every benefit of using AI in the UX design process, there are risks and pitfalls to avoid.

Speed, cost and quality

Advantages: AI can speed up UX design in numerous ways, from research and ideation to prototyping and testing. For example, it can accelerate concept iteration in the early stages of the design process compared to traditional UX methods. This has obvious implications in terms of reducing delivery timeframes and project costs.

Drawbacks: AI generated design ideas can be homogeneous, generic and lacking in consistency, meaning the time and cost of human input must be considered to ensure the final designs are sufficiently distinct and cohesive. (That is, if we care about designing better — or simply good — solutions, and not just faster solutions.)

Efficiency versus innovation

Advantages: The use of AI can relieve UX designers from mundane and tedious tasks, allowing them to concentrate on activities that require more critical thinking and emotional engagement. Essentially, there are huge opportunities to offload a lot of grunt work to AI systems. But designers need to be careful about this, because…

Drawbacks: Over-reliance on AI designs could lead to fixation on minor optimisations rather than out-of-the-box thinking. This means designers need to find a balance between efficiency and innovation — by avoiding dependence on AI where human creativity and agency would add value leading to better solutions.

Skills and development

Advantages: There’s clear potential for AI to lower the skill threshold required for designers in UX. Which makes sense, as prompting is easier to learn than all the features and functionality of Figma. This change allows a wider range of people to contribute design ideas without traditional visual design tool skills.

Drawbacks: Over-dependence on generative AI tools might impede development for UX novices, as repetitive tasks help cultivate UX design skills and judgement. Junior designers might be at higher risk of negative cognitive effects from AI generally, as young people exhibit higher dependence on AI tools and lower critical thinking skills.

Summary

It’s a mixed bag, basically. For all the opportunities that AI brings to UX design, there are plenty of challenges and traps practitioners can fall into. Some of these issues will be solved by technology improving, but others require skilful and mindful integration of AI into UX design processes.

3. How UX practitioners feel about using AI

These are reflections from UX professionals about their experiences of using AI in the design process. I’ve categorised them as either positive or negative sentiments, adding a few comments of my own (in parentheses).

Positives

• Using AI in the design process can make UX practitioners feel both more effective and efficient. (I hesitate to use the phrase ‘super-powered’, as The Incredibles tells us: when everyone’s super, no-one is.)
• Designing with AI helps develop skills in prompting, which is emerging as a core design skill. (It’s surely only a matter of time before we start seeing Prompting is Designing UX books hitting the shelves.)
• Generating design variants with AI avoids the ‘blank page’ problem, where UX designers struggle to get started. (Even for professionals, the blank canvas of a design file can sometimes be intimidating.)
• Collaboration with AI can feel like a complementary partnership between human cognition and AI technology. (Similar to a senior designer guiding a junior designer.)
• Using AI helps facilitate collaboration with stakeholders by streamlining idea sharing and exploration. (Basically, AI can help make the sometimes arcane practice of UX clearer and more accessible to stakeholders.)

Negatives

• Designing with AI can also feel like being a client commissioning a contractor rather than designing. (In other words: designers can feel like the AI is doing the fun, creative bits while they write the brief.)
• When UX designers feel like they’ve outsourced creative tasks, it can lead to diminished sense of ownership over the results. (How invested can you be if you didn’t even draw a single rectangle?)
• Creating an effective prompt can be a time-intensive process with a high cognitive burden. (Although in future this could be mitigated by refining prompting templates and support resources.)

Summary

The positives outweigh the negatives, but there are clear issues to resolve. This includes UX practitioners’ sense of identity as designers, as well as the time and mental effort required to craft effective prompts, generate outputs and carry out continual refinements.

4. Takeaways for future use of AI in UX design

I’ve tried to categorise these suggestions into meaningful themes. They typically crop up in the discussion and conclusion sections of academic papers.

AI increases efficiency, but people still matter

The biggest impact of AI in UX design is the increased efficiency across the design process. However, AI cannot replace human interpersonal communication, collaboration, creativity or originality; therefore it’s important to balance pursuing efficiency while preserving the human-centred nature of UX design.

Humans need to remain in the loop

UX designers should adopt a human-in-the-loop approach to validate AI output, improve model performance and avoid overreliance on automated systems. However, designers should be mindful that AI tools could potentially reinforce their existing biases instead of challenging them. This is an area where critical thinking skills are absolutely, well, critical.

AI policies are needed for ethical practice

Adoption of AI must be ethical and inclusive as well as efficient. To alleviate concerns about ethics, data privacy, ownership, and accountability, organisations should set up and communicate policies on AI usage. Too many UX practitioners are still driving ahead, often alone, without clear internal strategies or policies for generative AI.

UX designers need specific AI training

To successfully integrate AI into design team processes and workflows, UX practitioners would benefit from training to develop proficiency in crafting and refining prompts, assessing and criticising AI-generated output, and accounting for AI intricacies and limitations. There’s a world of difference between dabbling with AI in design and using it as effectively as possible.

Conclusion

So what’s the story? How is AI impacting UX design?

Well, it’s clear from the academic studies that traditional design methods are being innovated: UX practitioners are harnessing AI, leading to efficiencies and lowering costs for organisations in the process. The benefits are undeniable. UX design has fundamentally changed in the past 2–3 years, and it’s probably career suicide to deny that.

In fact, more UX designers might need to let go of conventional design stages and activities to allow for even more innovative processes empowered by AI. For example, strictly adhering to established workflows and handover processes ignores the potential to go from concept to functional, testable solution rapidly with AI.

However, there are potential drawbacks if AI usage in UX design is over-relied on, and used mindlessly. Without sufficient critical thinking, we can easily end up with generic, biased designs that don’t actually solve user problems. In some cases, we might even spend too much time on prompting and vibing with AI when we could have simply sketched or prototyped something ourselves — creating more sense of ownership in the process.

Many of these findings might feel obvious to a lot of people. But of course the point of rigorous research is often to validate what you think you already know. This deep dive into academia shows us what we really do know, for sure. For now.

Sources

Ahmed, A. and Imran, A. S. (2025) ‘The role of large language models in UI/UX design: a systematic literature review’, arXiv. Available at: https://doi.org/10.48550/arXiv.2507.04469

Batch, A., et al. (2023) ‘uxSense: supporting user experience analysis with visualization and computer vision’, IEEE Transactions on Visualization and Computer Graphics, 30(7), pp. 3841–3856. Available at: https://doi.org/10.1109/tvcg.2023.3241581

Bilgram, V. and Laarmann, F. (2023) ‘Accelerating innovation with generative AI: AI-augmented digital prototyping and innovation methods’, IEEE Engineering Management Review, 51(2), pp. 18–25. Available at: https://doi.org/10.1109/emr.2023.3272799

Buschek, D., Anlauff, C. and Lachner, F. (2020) ‘Paper2Wire: a case study of user-centred development of machine learning tools for UX designers’, In Proceedings of Mensch und Computer 2020 (MuC ‘20), pp. 33–41. Available at: https://doi.org/10.1145/3404983.3405506

Ebel, P., Brokhausen, F. and Vogelsang, A. (2020) ‘The role and potentials of field user interaction data in the automotive UX development lifecycle: an industry perspective’, In 12th International Conference on Automotive User Interfaces and Interactive Vehicular Applications (AutomotiveUI ‘20), pp. 141–150. Available at: https://doi.org/10.1145/3409120.3410638

Fan, M., et al. (2022) ‘Human-AI collaboration for UX evaluation: effects of explanation and synchronization’, Proceedings of the ACM on Human-Computer Interaction, 6(CSCW1), pp. 1–32. Available at: https://doi.org/10.1145/3512943

Gerlich, M. (2025) ‘AI tools in society: impacts on cognitive offloading and the future of critical thinking’, Societies, 15(1), 6. Available at: https://doi.org/10.3390/soc15010006 (Accessed: 12 September 2025).

Jiang, E., et al. (2022). ‘PromptMaker: prompt-based prototyping with large language models’, CHI Conference on Human Factors in Computing Systems Extended Abstracts, 35, pp. 1–8. Available at: https://doi.org/10.1145/3491101.3503564

Li, J., et al. (2024) ‘User experience design professionals’ perceptions of generative artificial intelligence’, In Proceedings of the 2024 CHI Conference on Human Factors in Computing Systems (CHI ‘24), 381, pp. 1–18. Available at: https://doi.org/10.1145/3613904.3642114

Luo, Y. (2025) ‘Designing with AI: a systematic literature review on the use, development, and perception of AI‐enabled UX design tools’, Advances in Human-Computer Interaction, 2025(1). Available at: https://doi.org/10.1155/ahci/3869207

Maceli, M., Smith, N. and Bhakta, G. (2024) ‘Incorporating unanticipated uses of generative AI into HCI education’, In Proceedings of the 6th Annual Symposium on HCI Education (EduCHI ‘24), 5, pp. 1–7. Available at: https://doi.org/10.1145/3658619.3658628

Padmasiri, P., et al. (2023) ‘AI-driven user experience design: exploring innovations and challenges in delivering tailored user experiences’, 2023 8th International Conference on Information Technology Research (ICITR), Colombo, Sri Lanka, pp. 1–6. Available at: https://doi.org/10.1109/ICITR61062.2023.10382802

Takaffoli, M., Li, S. and Mäkelä, V. (2024) ‘Generative AI in user experience design and research: how do UX practitioners, teams, and companies use genAI in industry?’, Designing Interactive Systems Conference, pp. 1579–1593. Available at: https://doi.org/10.1145/3643834.3660720

Uusitalo, S., et al. (2024) ‘“Clay to play with”: generative AI tools in UX and industrial design practice’, Designing Interactive Systems Conference, 9, pp. 1566–1578. Available at: https://doi.org/10.1145/3643834.3661624

Wang, Z., et al. (2024) ‘Exploring the impact of artificial intelligence-generated content (AIGC) tools on social dynamics in UX collaboration’, Designing Interactive Systems Conference, pp. 1594–1606. Available at: https://doi.org/10.1145/3643834.3660703

York, E. (2023) ‘Evaluating ChatGPT: generative AI in UX design and web development pedagogy’, In Proceedings of the 41st ACM International Conference on Design of Communication (SIGDOC ‘23), pp. 197–201. Available at: https://doi.org/10.1145/3615335.3623035

Zhu, Z., et al. (2024) ‘AI assistance in enterprise UX design workflows: enhancing design brief creation for designers’, Frontiers in Artificial Intelligence, 7. Available at: https://doi.org/10.3389/frai.2024.1404647

---

Silicon clay: how AI is reshaping UX design was originally published in UX Collective on Medium, where people are continuing the conversation by highlighting and responding to this story.

 

This post first appeared on Read More