Joyful ideation

The latest incarnation of GenAI is impressive for creatives. For example, Claude is highly regarded because it produces the widest cognitive range of ideas; Grok is a champion for "edgy" or counter-intuitive brainstorms; and Midjourney offers a complete visual exploration platform. Yet GenAI cannot replace human agency, physical presence and emotional intelligence. And while LLMs process data when prompted, only humans possess intrinsic curiosity, true creativity and ethical reasoning. And then there is the pleasure of ideating. Like airline pilots who love flying manually, despite all available technological support, so the best designer love designing without having to rely on AI, at least in the ideation phase. This is because people love doing what they are good at. A key word here, then, is the joy of of designing. Or, as the architect Peter Cook puts it: "architecture should be a cheerful business". That is, playful optimism translated into productive ideas.

GenAI-free workshop?

The ideation workshop was launched well before the mainstream use of GenAI, such as ChatGPT (2022) when designers generated ideas stemming from personal iteration with pen and paper, collage or sketch modelling using materials such as card- and foam-board, balsa wood, clay, wire or plastic sheets. These ideation tools are known for their immediacy, low-cost and flexibility supporting human creativity and traditional design principles. And when the analogue tools are extended by software tools such, as Adobe Photoshop and Rhino 3D, they combine physical and digital crafts. However, what has changed with AI is that machine learning models (LLMs) can simulate human ideation tools generating new content in form of text, images, and video. Yet there are differences. And so, whereas non-GenAI ideation tools rely on human intent, intuition and lived experience as well as technical skill, GenAI tools depend on user prompt, vast training data, high-performance computing and algorithmic pattern-matching. That is, GenAI tools are reactive, the prompt initiates the ideation path which is then shaped by the AI model. But the design process is supported by multimodal tools (analogue, digital and AI), and designers typically switch between tools or use them in different configurations informed by the complexity, constraints and scale of the design project. Ideation output, however, is evaluated not on the type of tools used but by human review and judgement. Yet for reason of clarity, transparency and integrity, all tools used in the ideation process should be openly declared as part of the idea(s) presentation.

Ideas thought through

Le Corbusier, the Swiss-French architectural designer, emphasised the role of ideas in the design process saying that nothing should be built without being "thought out" or "no action before thought". That is, rather than engaging in impulsive construction, Corbusier's philosophy tied physical form directly to underlying concepts such as his five principles of modern architecture introduced in 1926 (pilotis, the free floor plan, the free facade, horisontal ribbon windows, and roof gardens). But how does the notion of "thought out" manifest itself in the age of AI? Corbusier's intellectual position was guided by "unity of intent", or spirit of order which GenAI is lacking as it operates on a mathematical order to predict what looks or sounds right. And while GenAI output may feel "thought through", and appear brilliant in places it may not necessarily do so in critical "real-world tests". That is, GenAI's output represents statistical probabilities and doesn't "think" or understand concepts the way humans do. However, humans set the conditions (prompt commands) to which GenAI responds. This makes GenAI a useful design tool, for example, to overcome creative blocks, expand conceptual horizons or facilitate novel solutions. Indeed, used thoughtfully and skillfully, GenAI may help designers focus on empathy and while complex make the design process more human, not less. Corbusier, who was a proponent of new technologies beyond utilitarianism, would probably agree.

Thinking aloud with AI

AI technology, which includes AI chat interfaces and image generators, has pushed ways of ideation beyond traditional sketching on paper or digital drawing programs, such as SketchUp. Both modes are good at drawing at the speed of thought. But while visual elements are core to design, verbal language is the main means of human communication as exemplified by AI systems transforming speech into sketches. That is, by dictating, or thinking aloud to a Large Language Model (LLM), such as ChatGPT the program converts spoken ideas into visual sketches in real time. Or, by combining freehand drawing on screen with real-time speech input also known as sketch-and-speech fusion*. And so, GenAI enables ideation using speech input as alternative to text or visual prompts. AI technology, then, is modifying and expanding ways of generating and communicating ideas that, moreover, can help widening participation in the design process to those not familiar with conventional ideation tools such as pen and paper sketching. However, any conceptual tool, analog, digital or AI supported depends on the users' skills. * https://www.researchgate.net/publication/397480818_TalkSketch_Multimodal_Generative_AI_for_Real-time_Sketch_Ideation_with_Speech

AI boost to design possibility

The impact of AI technology on ideation is complex because the nature of AI is dynamic, like ideation itself. But also contradictory, as cursory observation or casual conversation in design settings, both in education and professional practice reveals, in particular disagreement over both the usefulness and transformative nature of GenAI. For example, given the vast size of data sets and processing power, GenAI suggests plausible-sounding outcomes of high certainty whereas what they actually produce are calculated (statistical) probabilities based on algorithms (lists of instructions). In contrast, ideation is not about probability or certainty but inspiring suggestions or creative propositions. But when GenAI is being used for ideation, their outcomes depend on how apt users are at prompting the generative models and this calls for reflective questioning and interpreting ambiguity and unique circumstances. This suggests that combing human agency, including traits such as critical thinking, intentional action, judgement and empathy with AI systems increases design possibility. That is, designers can explore new forms using traditional modelling software with AI features while keeping human oversight and control of the final results. In other words, analogue and digital worlds co-exist because they supplement and enhance each other.

GenAI: the Joker in the ideation pack

Design ideas can be described as either visual, concrete or abstract. But, in the age of artificial intelligence, AI, how do designers approach the ideation process and how do they generate and communicate their ideas? What ideation tools do they use, and do they show preferences for certain tools? While visual or verbal language is the main means of communicating ideas, that is, language as an ideation tool, the range of tools is wide, contextual and embracing both analogue and digital modes.. And so, for visual ideas, designers use traditional pen and paper sketching or physical or computational modelling. For concrete ideas, they explore physical forms and materials using craft methods or digital fabrication. They become curious about the theoretical underpinnings of design, as expounded by scholars such as Donald Schön, Victor Papanek, and Nigel Cross. They get inspired by design history or designer biography, for example, the inventor James Dyson. Then AI, which is greatly influencing, if not transforming the study, research and practice of design across disciplines. In this digital environment, GenAI has become a game changer that could be perceived as the Joker in the ideation pack (a trump, or a wild card). Or, the Magician, from the imagery of tarot cards (the card of new beginnings and opportunities). Finally, there are motivational sources that power new ideas and innovation: human reason, emotion or desire (although not necessarily separated states of mind, and not easily measured). Yet, not to forget  the ethical dimension of design in a world of conspicuous consumption and environmental concerns.

The human factor in ideation

It is now commonplace to start the ideation process by prompting generative AI tools, such as ChatGPT, rather than using traditional search engines, such as Google. And so, whereas Google is typically used for direct questions (Q&A), ChatGPT is used to generate text and images based on conversational interaction which makes the chatbot a handy and speedy tool for idea generation. Google, however, has responded to the chatbot challenge by offering AI functionality to its traditional search engine, aka "AI mode". Yet, depending on the context, a traditional search engine might work as well as, or even better than ChatGPT. That is, in many ways, search engines work with the same fundamental information as traditional search engines. This highlights the interaction between people and AI tools and the need to recognise the non-technological side of the design process, including privacy and ethical risks in AI projects. That is to say, the role of the human factor in innovation might be overlooked or undervalued in the AI matrix of speed and efficiency - an issue the workshop tries to address.