Skip down to main content

How AI is Changing the Design Process

How AI is Changing the Design Process

Published on
26 Jul 2023
Written by
Maggie Mustaklem
AI is changing how designers search for inspiration. What does this mean for the future of design asks OII researcher Maggie Mustaklem.

AI is changing how designers search for inspiration. What does this mean for the future of design asks OII researcher Maggie Mustaklem

Designers used to turn to books and magazines to search for inspiration. Now they primarily search online, using social media tools like Pinterest and Instagram. These social media platforms use “everyday AI”, including Natural Language Processing (NLP), machine learning and computer vision. Their use in design research may influence what is sketched, prototyped and ultimately produced. My doctoral research at the Oxford Internet Institute is investigating how this “everyday AI” is changing design inspiration.

Any form of AI in Use Warrants Critical Inquiry

AI is evaluated critically in many contexts, from facial recognition to election misinformation. If the platforms designers use for inspiration are the very same ones culpable of well-established biases in other areas, it stands to reason the design process also deserves our attention. It is more difficult to identify what may potentially be insidious about the distribution of photos of monstera plants, coffee shops or cool jumpers. But the design industry is huge, valued at $162 billion in 2021, and this is an applied context of AI in use that enacts upon the buildings, products and services we use worldwide.

What’s Missing in Digital Negative Space?

In art and design, negative space is the empty space around an image, the white background surrounding a black shape. When it comes to thinking about bias in the context of design inspiration, digital negative space is a helpful framework. There are billions of images online, but what don’t we see when we search online?

Take the contrasting examples of Dutch and Indian design on Pinterest. A search for Dutch design returns sleek contemporary design, not wooden shoes. On the other hand a search for Indian Design returns handcrafted traditional block printed textiles, saris and hand carved furniture. Modern Indian product designers exist, but it is much harder to find them online. Who gets to be modern in this context, and who gets to influence contemporary design? A fetishized, colonial version of India narrativizes what users of social media platforms see, and what they don’t. Chasing Innovation outlines how national identity and the politics of aesthetics affect the “Indianness” in the design process more broadly. My research focuses this query on what designers see when they are searching for inspiration. Dutch designers are renowned and among the most financially successful in the design industry. AI may be unwittingly reinforcing their success while marginalizing contemporary Indian designers, relegated to digital negative space.

dutch design

“Dutch Design” searched on Pinterest returns contemporary designs.

indian design

A search for “Indian Design” on Pinterest leads primarily to traditional block printed textiles.

Design. Interrupted

Understanding the social and cultural significance of digital negative space allows us to evaluate design inspiration as a practice mediated by AI-enabled social media platforms. Through Design Interrupted, part of my doctoral research project, I have been evaluating how AI is changing design inspiration for professional designers and architects. Over the course of fifteen workshops with creative studios in London and Berlin, I asked designers to reflect on their search practices through practical exercises. I had designers begin by leafing through magazines for inspiration to create analogue mood boards. This required designers to think about systems like Instagram and Pinterest as part of a collective search process and led to insights about what is missing and what could be improved with current tools. There are a range of methods designers can use to engineer more varied, serendipitous searches online. For example, instead of starting with a final product search term, breaking a search down into components that need to be collaged (by a human) may lead to less prescriptive results.

What about Generative AI?

In the real world, design research operates on compressed timelines, making it a task suited to further automation. But as one participant said, “this is the fun bit.” Designers get a lot out of doing research. Beyond product deliverables, design research expands designers’ knowledge in ways that may inform future work. Most studios are toying with generative AI but have yet to replace existing workflows. That will likely change at some point in the near future.

As the relevance of tools like Midjourney and DALLE-2 grows, they ought to be seen as more of a gradual evolution of existing systems. There are numerous stories emerging of bias in the training data for Large Language Models (LLMs), mirroring existing criticisms of AI tools currently in use.  Additionally, LLMs are primarily text-to-image. This requires a professional’s situated, embodied understanding of the prompts they use. This is precisely the kind of skill designers develop by doing their own research. As this space evolves, it is important to critically reflect on the role played by aesthetic images shared online in the design of products, buildings and experiences that shape our world.

If you work on a creative platform or for a design firm and would like to contribute to research on the topic, please feel free to reach out to

Find out more about Maggie’s research.

Related Topics