Learning from Frank Gehry

The buildings by the architect Frank Gehry (1929-2025), such as Bilbao’s signature landmark, the Guggenheim Museum (1997), has gained wide public admiration as well as inspiring many designers. But equally his design philosophy: 'To design something that one would want to be a part of, something one would want to visit and enjoy in an attempt to improve one’s quality of life.' More specifically, Gehry considered architecture 'to be art' aiming at transfering the feelings of humanity through inert materials. Not surprisingly, he greatly appreciated sculpture which influenced his architectural approach. He has been known for his deconstructive approach and structural choices using non-rectilinear shapes and forms. In this pursuit, right from the beginning, his work experimented with rough, even industrial materials. But his process began with freehand sketches and models visualising what he had in mind, then realising his ideas using advanced architectural 3D modelling software. Gehry also took a keen interest in education encouraging students to always be curious, and let architecture open up to other subjects, such as philosophy, literature, and music. He also advised students to study and learn from the greats, for example, Le Corbusier, Zaha Hadid, Brunelleschi, Oscar Niemeyer, Louis Kahn, Ludwig Mies van der Rohe, Eliel Saarinen, Louis Sullivan, and Frank Lloyd Wright.

AIdeation: ChatGPT three years on

ChatGPT, the generative artificial intelligence chatbot, was launched three years ago (November 2022) and has evolved rapidly to become the go-to software tool for generating, developing and communicating ideas, enhancing concepts and exploring scenarios. But more than this, ChatGPT, together with CAD and 3D printing is increasingly facilitating and streamlining the design process - from first thoughts to prototyping. Indeed, rare is the designer who hasn't engaged with ChatGPT, which holds 61 percent of the market for Generative AI chatbots. But despite its widespread adoption, ChatGPT is not without risks. For example, it can pose a threat to integrity, such as plagiarism (although copyright does not protect ideas or concepts per se), or create technology dependency. But while the impact and implications of ChatGPT are felt and experienced across design fields, both in education and practice, the chatbot has become a powerful assistive design tool offering creative synergy between human ingenuity and AI. Yet like any computer tool, ChatGPT should be used from critical and informed perspectives, considering both its benefits and limitations as a creative source for best course of action or possible outcome. After all, ChatGPT is only as good as the user's prompts. This highlights the role of judgement in creative thinking which may suggest building foundation for critical thinking without relying too much on GenAI systems, especially for novice designers. That is, to encourage critical thinking, in the context of design education and creative learning, to develop fortuitously through personal engagement with tools and materials that offer tactile sensations and emotional connections. That is, to empower users of ChatGPT, in often cluttered and noisy digital environments to (re)discover arts and craft tools for inventiveness. But overall, the challenges presented by chatbots call for ongoing experimentation, research and discussion.

Ideation, GenAI and critical thinking

In an analogue world, design ideas aren't forced upon the designer. They are conjectural, or guesses about reality - a proposal or tentative solution to a posed problem. In this pursuit, designers aren't idealists because design ideation is a purposeful activity aiming at realisable ideas. The designer, then, is seen as a realist seeking to be proven right. That is, the designer faces practicality having to accept the physical facts of the situation and, oftentimes the emotional side to the problem at hand. Yet some ideas can clash with reality, and when they do they remind designers that ideas may be mistaken. Ideation, then, is a process that must allow criticism in order for the idea to move forward, to propose a better solution. That is, ideation includes critical thinking skills. And so, designers not only need to be imaginative and open-minded but willing to be corrected. However, with the rise of Generative AI, what is its impact on critical thinking? Interestingly, a recent survey (2025) shows that in GenAI-assisted tasks higher confidence in GenAI is associated with less critical thinking, while higher self-confidence is associated with more critical thinking.* This may suggest that practsing ideation without GenAI assistance could help foster greater critical thinking skills also raising designer self-confidence in problem-solving ability. https://www.microsoft.com/en-us/research/publication/the-impact-of-generative-ai-on-critical-thinking-self-reported-reductions-in-cognitive-effort-and-confidence-effects-from-a-survey-of-knowledge-workers/

Ideation and problem-solving

When designers respond to a problem, the problem is typically on a sliding scale from simple to complex.  Roughly speaking, simple problems can be described as puzzles, or "tame problems", whereas complex problems are known as "wicked problems". A puzzle is fairly straightforward when the pieces need to be located and connected. That is, it is assumed that the solution to the puzzle is almost certain to be found. A complex problem, in contrast, has no known or at least immediate solution and so designers are faced with challenges that call for a greater variety of skills and capabilities involving collaborative and decision-making tools and techniques. But whether simple or complex, the starting point for problem solving includes clarifying the goal, identifying the constraints, and understanding the context of the situation. Also, see Tame vs Wicked problems, in blog below. https://www.td.org/content/atd-blog/puzzle-problem-challenge-or-conundrum

Conversational prompting

Prompt engineering is the process of writing effective instructions for a GPT model, such that it generates content that meets desired outcomes. And because the content generated from a model is non-deterministic, that is, even for the same input, the model can generate different responses on different runs - and the rich datasets combine art and science, prompting can be particularly useful in creative fields such as design. However, there’s more to effective human-AI collaboration than a perfect prompt. And so, conversational prompting is a technique that involves interacting with AI systems like ChatGPT in a human-like conversation. That is, users describe in everyday terms what they want ChatGPT to do rather than trying to craft complex prompts. In this, users engage in back-and-forth interactions with the model to refine the results, provide additional context, and answer the ChatGPT's questions. That is, the user guides the process while letting ChatGPT handle the specifics of generating appropriate prompts and responses. Problem solving, then, it is thought, is enhanced by balancing human ingenuity and machine intelligence. However, conversational prompting, as a feed-back model, carries risks too. That is, the interaction may contain misinformation, biases or illusions raising the question: Is the output reliable and trustworthy? Selected sources:  https://platform.openai.com/docs/guides/prompt-engineering    https://promptengineering.org/conversational-prompting-in-generative-ai/

Tame vs Wicked problems

Many design problems have fairly straightforward solutions without needing a particularly ‘creative’ approach – these are often technical in nature and often described as tame problems. In contrast, wicked problems are complex sets of problems and even when of a technical nature difficult to define with no straightforward or obvious solution. However, some problems have properties of both tame and wicked problems and some problems are neither wicked nor tame but remain in a state of uncertainty. Yet the distinction between tameness and wickedness resonates with the difference between traditional AI, such as Google's search algorithm, and generative AI, such as ChatGPT. The main difference between traditional AI and generative AI lies in their capabilities and application. That is, traditional AI systems are typically designed with a specific set of tasks in mind, primarily used to analyse data and make predictions based on predefined rules, while generative AI goes a step further by creating new data similar to its training data. In other words, the strength of traditional AI is pattern recognition, while generative AI excels at pattern creation. This suggests using GenAI to help dealing with wicked problems which are impossible to solve in a way that is simple or final. However, the uncertainty or ambiguity surrounding wicked problems requires human evaluation of any problem resolution. That is, AI systems are tools, not sentient beings. https://www.forbes.com/sites/bernardmarr/2023/07/24/the-difference-between-generative-ai-and-traditional-ai-an-easy-explanation-for-anyone/

Ideas to market

Generative AI has changed, and is changing design ideation. Today, it is nearly impossible to avoid AI-powered search engines, such as ChatGPT when generating ideas. Indeed, AI is quickly changing how people search and use the internet more generally. Instead of getting a set of links to follow from conventional search engines, such as Google, AI chatbots provide user queries with short direct answers. And so, the latest GenAI chatbots can produce text, image or sound content without the need for human intervention. Arguably, however, few of AI generated ideas are truly unique or original as GenAI outputs are derived from the input data used to train the AI tool. Yet there are limitations to human creativity too. Indeed constraints present challenges and foster innovation. Therefore the held view that GenAI, as an ideation tool augments human creativity. But whether generated by humans or AI, ideas made public are not protected by copyright unless the idea copied is an expression of a specific work. This suggests that designers who seek to exploit their ideas sufficiently express them to meet the criteria for copyright. Moreover, in a competitive market, the idea need not only be expressed as a specific work but also brought to market, and fast. However, having turned the idea into a product or service, and made it available to the market before anyone else doesn't guarantee commercial success.