Ideation and reflective design

The application of generative AI, GAI to design ideation raises the question what is the imapact of GAI on reflective design. That is, designers typically reflect on their actions, both during the design process ("thinking on one's feet") and after it, in what Schön calls reflection-in-action and reflection-on-action (The Reflective Practitioner, 1984). Moreover, reflective design, which implies critically thinking about the design process, suggests that designers need to reflect on the unconscious ways in how technology, such as GAI shape designers'  decision making, actions, and design proposals. For example, what is the impact of GAI, such as communication with a text-to-text model or text-to-image model, on designer ability to uncover hidden biases or limitations on design practice. In Reflective Design (Sengers et al, 2005) the authors argue, when designing with technology: 'Each person has preconceived notions shaped by their personal journey and that affects their practice as a designer. Thus, one needs to unpack personal ideologies and determine how they influence the reflective design processes.' Moreover, the authors state: 'Technology shouldn’t be viewed as the optimal medium for design; technology imposes its own set of constraints and values to its users which should be addressed.' The notion of the reflective practitioner, and reflective design, then, serve as a guide for designers in the age of AI too.

Solo or group ideation?

Design is typically a collaborative activity, at least for major projects that involve many stakeholders. But what about the early stages of the design process, in the phase of generating ideas? Does ideating on one's own result in substantially different ideas compared with ideating in a group (typically five or more people), including so called "brainstorming"? Research suggests that in the workplace people become less productive in a team compared to working alone. This phenomenon has been labelled "social loafing" whereby people are said to be less likely to feel personally accountable for the outcome of a task when they are working in group. So would the concept of social loafing apply to ideation? In other words, is there a change in motivation among ideators when the ideation task is shared in group work? This, of course, would depend on the task at hand, or the composition or dynamics of the group. Moreover, working together can motivate people to perform well but it can also lead to a loss of motivation because the individual contribution is not as visible. As in "who came up with the idea first, and who can take credit for it?". So ideating in group may be a mixed blessing. A way out of this dilemma would be to start ideating alone followed by discussing individual outcomes with the group.

The power of sound

Ideators typically generate ideas through words and images: "In the beginning was the word" or, "a picture says more than a thousand words". Or the dialogue between language (verbal) and visualisation (non-verbal), also described as the picture-word-cycle. Interestingly, the architect Frank Lloyd Wright, in 1930, worked on his lectures for Princeton University: 'I must say I have enjoyed writing them'. The six lectures didn't need images, he added. 'Personally, I've never cared much for illustrated lectures. If we might have some good music to begin with and end with - that might help'. Yet Wright's writing and public lecture series provided a new outlet for his ideas. Moreover, on the power of music, the pianist and conductor Daniel Barenboim, in the first of his five BBC Reith Lectures, in 2013, argued, in examining the transformative power of music in the world: 'In the beginning, there was silence. And out of the silence came the sound'. Furthermore, it has been said that in the beginning was sound. Sound began the whole thing, and in sound resides tremendous power. It opens doorways to other realities, for with the production of sound, energy can move from one system to another. Now, there's a thought for ideators: The power of sound stirring imagination and innovation. Source: Wright quote: Alofsin, A (2019) Wright and New York. New Haven: Yale University Press. Barenboim: https://danielbarenboim.com/the-phenomenon-of-sound/

GenAI feeds ideation

Generative artificial intelligence, GenAI is already being employed as a tool by designers in what is referred to as machine co-creativity and augmented design, with the technology not only facilitating ideation but also used as a visioning tool (text-to-image software) for community engagement for public project. But there are limits to GenAI applications, notably issues of context, quality and data bias. That is, GenAI ‘learns’ by analysing existing data so it is afflicted by evidence of in-built bias - which highlights the risk of, say, relying on GenAI for a housing concept. That is, while GenAI is useful as a tool for designers to help create an architectural image, or to generate a seamless to-do-list for housing developers, it is not a blueprint for a complete building scheme, notably for aspects of detailing or from a structural engineering perspective, for which BIM (Building Information Modelling) is typically used as interface for a unified design environment.

Unmeasurable ideas

'A great building', wrote the architect Louis Kahn (1901-1974), 'must begin with the unmeasureable, go through measurable means when it is being designed, and in the end must be unmeasurable'. The same thinking, arguably, may be applied to ideation. That is, ideas that spring, or emerge out of dreams or relatedness. And whether the idea takes centre stage or appears in the margins, first thoughts seem unmeasurable. But, in the development of an idea, measurable means are employed to make the idea visible or comprenhensible, and this in order to express and communicate the idea, visually or verbally. And only as far as the idea is made real, that is, realised, does it become measurable. But not necessarily so because ideas often carry intangible qualities in which case the idea, in the end, must remain unmeasurable. Could it be then that the measureable idea begins where dreams end?

AI as co-design partner

Co-design, or participatory design is an approach to design which focuses on processes and procedures of design. It has been used in many settings and at various scales. Rooted in work with trade unions in the Nordic countries, this approach has also a political dimension of user empowerment and democratisation. From this persepctive, co-design requires the end user's participation: not only in decision making but also in idea generation. Moreover, it is argued that designers create more innovative ideas and concepts when working within a co-design environment with others than they do when creating ideas on their own. Interestingly, AI has now entered co-design environments through large language models, LLMs, such as chatbots, which are a type of artificial intelligence algorithm that uses deep learning techniques and masive data sets to understand, summarise, generate and predict new content. LLMs, then, are able to process vast amounts of text data, mostly harvested from the Internet providing a basis to generate and communicate new ideas and concepts. Moreover, LLMs use natural language to simulate human conversation.The model, then, as a conversational chatbot becomes effectively a co-design partner.

Synthetic ideation

Synthetic media is a catch-all term to describe video, image, text, or voice that has been fully or partially generated using artificial intelligence. So, synthetic media includes ideas that have been fully or partially generated and presented using generative artificial intelligence, GAI. Conversely, non-synthetic media can be any other media, that is, media that is produced by human input. Take a freehand idea sketch, for example. Drawn entirely by a human, the rough sketch would fall under non-synthetic media, or non-synthetic ideation. Now think of a human-generated freehand sketch which has been digitised and where an GAI image editor is used to add a feature to the sketch. This new (altered) image is partially generated by AI (the picture being modified by algorithms), and we can thus say the idea, as represented by the image is “synthetic.” Ideas thus generated, or synthetic ideation may pose the question; "who is better at ideation - GAI or human?" GAI has brought a revolution in image editing, that is, a technology that turns manual image editing processes into automtic or semi-automatic actions. GAI, then, in drawing on a vast and varied image bank (datasets), and at speed, is a powerful ideation tool. Yet GAI is not a simple substitute for human decision-making and to fully appreciate the idea, the discerning eye (judgement) will be human, at least for now.