Food for thought: Ultra-Processed Design?

A diet of ultra-process food, UPF, when consumed in excess is generally thought of as having negative health implications, such as the increased risk of obesity, diabetes or anxiety. But while processing food is not necessarily a bad thing, what makes ultra-processed foods distinctive is that they have gone through industrial processes that have changed the nature of the original ingredients, leaving little, if any, of the original whole food behind. Could such concerns apply to the design process mediated by artificial intelligence, AI? That is, AI data as ultra-processed food. What makes AI distinctive is that it combines various algorithms to represent and process content. For example, to generate text, various processing techniques transform raw characters (e.g., letters, punctuation and words) into sentences, parts of speech, entities and actions which are represented as vectors. Similarly, images are transformed into various visual elements, also expressed as vectors. But is it meaningful, or fair to compare AI data with UPF as it may conjure up images of AI junk, AI addiction or AI obesity. Dietitians argue that it is  the overall dietary pattern that matters, not individual foods. Sufficient to say, then, that it is the mix of design tools that matters, not AI per se. Yet the comparison can be informative and contribute to healthy conversations about the role of AI in design.

Risk of over-dependence on AI?

Skills that once were core to designing, such as hand drawing and rendering, have become less prevalent and less practised in the digital age. And as computers perform many tasks that once depended on designers independently, artificial intelligence is impacting the design process from first thoughts to final outcomes. But as the use of AI among designers has become widespread, reliance on algorithmic systems designers might become too dependent on AI-generated content that, moreover, might be too similar to each other. That is, given the same prompt by various users, GenAI generates the same output. This risks a feedback loop where AI, trained on AI-generated data results in diminishing returns of new ideas blurring the line between innovator and imitator.*. The counter argument is that AI simply enhances the creative process through collaboration between machines and humans? That is, combining AI with human intuition and emotional depth might yield outcomes that neither AI nor humans could achieve alone. Anticipating AI evolution, then, is difficult as algorithm development is fluid and dynamic (fluid non-linear dynamics are utilised in machine learning) and with each new development, new potential risks and opportunities arise. The challenge for designers, then, when ideation itself reflects transition and transformation, is to leverage GenAI to facilitate and enrich the design process.  * 58% of British architects say AI increases the risk of their work being imitated, in RIBA AI Report 2024.

Complexity and contradiction in ideation

Design ideas, pragmatically speaking can be characterised as either simple or complex where the simple idea proposes a fairly straightforward, low risk solution to the problem at hand whereas the complex idea combines several simple ideas to produce one idea but at a higher level of uncertainty. Designers deal with both simple and complex ideas although simple ideas are typically easier to generate and communicate to clients and other stakeholders. Overall, clients prefer simple ideas as they are easier to understand and execute and are less likely to go wrong. Yet complexity reflects the multi-layered reality of the world around us which calls for deeper levels of understanding and greater problem solving ability on the part of designers. So what about contradiction between simplicity and complexity? Design ideas, as elements of thought, overlap and interrelate and are developed and articulated through the project. And so, ideation can be seen as a puzzle of simple and complex ideas put together by the designers and stakeholders involved in the project. In this pursuit, Generative AI has become a powerful ideation tool, and a virtual member of the design team.

2025: What have we here?

In a famous 1903 letter, the Austrian poet Rainer Maria Rilke (1875-1926) urges a protege of his to 'be patient toward all that is unsolved ... to try to love the questions themselves like locked rooms and like books that are written in a very foreign tongue … The point is, to live everything. Live the questions now.' And so, what have we here? A bit of self-help that suggests how questions generate ideas, and vice versa. In other words, questions trigger ideas that pave the way for design solutions.