Artificial Intelligence design, an augmented creation process
Artwork by Studio Tim Fu
The present and future of design and AI formed the subject of one of the Round Tables at the Salone del Mobile 2024. Their relationship can be a highly productive one, but it’s worth investigating the modes, times and possibilities to get a grip on the changes taking place
Artificial Intelligence and design was the subject for discussion between Paolo Ciuccarelli (Center for Design at Northeastern University and DensityDesign Lab at Milan Polytechnic University), Sherif Tarabishy (Foster + Partners), Hannes Kochm and Florian Ortkrass (Random International) and Cas Esbach (MVRDV), moderated by Ed Stocker (Monocle), at a Round Table organised as part of the Salone del Mobile di Milano 2024 cultural programme. The Present and Future of Artificial Intelligence was an opportunity to tackle the subject of the deep-rooted relationship between creative activities and computational capacity.
Aside from the fears expressed around the possibility of this particular tool taking over designers’ work to a more or less massive extent, what emerged from the discussion was a revolution with regard to the relationship with the tool and the possibility of setting up a prolific discussion with it. This last point substantially changes the possibilities offered by Artificial Intelligence.
Being able to structure a relationship with a tool lays the foundations for an exploration that, theoretically, can go in an infinite number of directions. This characteristic means that it is not comparable to CAD or to any other tools utilised by designers. Unlike the latter, in the relationship with AI it is no longer just the designer who bends and adapts to the tool but, like a pas de deux, both the tool and the user bend and adapt to achieve a common goal. If this constitutes a turning point with regard to the tool, let us try to understand how this reciprocal shaping comes about.
Artificial Intelligence and the design process
A diagram called the Double Diamond was first devised in the midst of the neo-positivist wave during the early 2000s, intended by its creators to describe the process of design and innovation. The design process continues to be a complex system, made up of convergent and divergent phases, in which creative thinking goes head to head with the possibilities and limitations that design itself creates for designers and companies. The limits and the perimeter of action are given by various parameters that may be dictated by the materials, the budget, the users’ expectations and market data. Today, in the age of Artificial Intelligence, that diagram is flanked by a new one that considers the synthetic tool as an integral part of the design process. The Stingray model splits the design process into three fundamental parts, taking into consideration the continual involvement of Artificial Intelligence. The macro-stages into which the new design process can be split, harnessing AI, are Train, Develop and Integrate. On the basis of this snapshot of the process, how can Artificial Intelligence alter and therefore impact design from a product/service point of view and from the point of view of exquisitely human activities?
Training Artificial Intelligence
At the beginning of every design process there is an ideation phase. This is the stage of the design process most closely linked to human imaginative capacities. A phase in which the knowledge, experience and sensitivity of the designer play an important part in shaping the brief and the theme of the work. In this phase, inspiration, reference, personal taste and design instinct combine to achieve the goal of conceiving something unique and desirable. Here, then, is the first stage at which the capabilities of AI come to support the designer's activity, in a phase in which the human presence guides and trains Artificial Intelligence. By utilising its abilities to connect and to process data, the designer’s creative activity can be boosted in a process of Augmented Creativity. A number of different algorithms can be used during this phase: from the most famous ChatGPT which can help designers to hone their ideas, find inspiration, gather and process data and even refine their concepts, right down to the tools that use generative algorithms to create images, as in the case of Stable Diffusion and Midjourney. Some examples of this are the chairs designed by Philipp Schmitt & Steffen Weiss, previously mentioned in this article, and the 20 sustainable packaging solutions that the New York-based BOI consultancy firm designed in tandem with AI.
AI has also been employed in the architectural field, for the new Autodesk building in Toronto for example, for which more than 10,000 layouts for optimising the space were explored.
Process development and optimisation
The ideation phase is followed by the development phase, during which Artificial Intelligence supports designers in their formal choices in order to optimise the production processes, increase resistance or the physical/material features and minimise waste and use of materials.
The most interesting examples of this include the optimisation of the Airbus A320neo, planes that, thanks to the use of AI at the design stage, are considerably more fuel efficient when in flight. Again in the field of aeronautics, DJI, an industry leader in the design and marketing of drones, uses AI to design drones with better flight, stability and autonomy. AI also helps to optimise the shape and weight of drones for a range of applications, such as aerial photography and industrial inspections. It was also harnessed in the design of the Elbo chair by Autodesk, in tandem with a generative algorithm, based on various simple parameters: a style inspired by Nordic modernism, a height of 46cm from the ground and weight-bearing to 136 kg, the machine did the rest.
Iteration and designer training
The final phase of the design process involves a lengthy iterative stage with the experience and training of the designer informing the use of the synthetic tool. In this phase, the designer takes on the role of the curator, using AI to produce tests and refine the design choices selected thus far.
From this point onwards, the use of AI could evolve not just as an ideational tool, but also be incorporated into products, in order to make changes to socio-cultural paradigms by changing habits and living spaces.
Samsung has done this in many of its new appliances, such as the Family Hub collection of refrigerators. These smart fridges are equipped with touchscreens and use AI to manage food, suggest recipes based on available ingredients, create shopping lists, and track food product expiry dates. In addition, virtual assistants such as Bixby can be integrated for voice control.
So if the question is "Will AI replace designers?" the answer can only be no, but what it will do is profoundly alter their work by increasing its possibilities, modifying some steps and enhancing some of its aspects. AI is therefore a powerful tool that, due to its characteristics, does not simply slot into the design process, but alters and changes it, requiring a continuous relationship of forces and adaptation from designers which could yield extremely interesting results.
