Even seen from a distance, the AI chair Philippe Starck has made with Artificial Intelligence (produced by Kartell, and presented at the Salone del Mobile) still looks precisely like a chair by Philippe Starck. Yet at the press conference he insisted: “the computer did it, I didn’t.”
It comes naturally to wonder where the advantage lies in a project done by a computer, though still with the guidance of a designer, if the final result totally resembles what the design would have done on his own. Apart from the communicative impact of the operation, the real benefit is productive efficiency.
This is nothing new. Already some years ago, the multi-disciplinary think tank Pier 9 in San Francisco, owned by Autodesk, the company that provided the software for Starck’s AI, produced a small object that seemed to be made of bone, but actually represented the best possible connection between the handlebars and the frame of a bicycle. An object generated (and then 3D printed) not by man, but by a computer. Although the designer had set the objectives and the fundamental parameters in which to operate.
Generative design, in fact, represents a process in which the designer communicates with software (Starck himself talks about a ‘dialogue’ with the machine), giving it precise orders. For example, the order to design the best connection between the frame and the handlebars, based on precise requirements: to guarantee a given level of shock absorption, to comply with preset minimum and maximum sizes, to use the smallest possible quantity of material, to achieve the lowest price possible.
It is amusing to imagine a printer that produces piles and piles of sketches: actually, everything happens in a paperless environment, but there are hundreds of successive projects. The designer (in flesh and blood) has the job of selecting the best ones, and perhaps perfecting them.
In the field of aeronautics, generative design – i.e. artificial intelligence applied to design – has been practiced for years. In 2016, in fact, a ‘bionic partition’ was presented for the Airbus, developed by The Living, a studio for Autodesk experimentation based in New York that applies data from the biological world to generative design projects. Perforated but stronger than the partitions of the present, thanks to a design based on an algorithm that imitates the growth of human bones, it permits 55% reduction in weight, which means much greener airplanes.
Airbus decided to go forward with research in that direction, inking a pact with the multidisciplinary think tank 3DExperience of Dassault Systèmes, which already stood out for the invention of Catia, the parametric design software – a precursor of generative design – used by Frank Gehry for his architecture, starting with the Peix d’Or in Barcelona in 1992.
“Starck was the one to suggest this experiment,” says Lorenza Luti of Kartell. “Philippe was in contact with Autodesk, which had the technical know-how to approach such a project, but had never worked in the world of furnishings. It is true that the final product inevitably ‘looks like Starck,’ but this is because he directed the development process: Philippe provided the algorithm with the main inputs, insisting on the need to use as little material as possible and then refining the formal output. Nevertheless, the angles, radii and curvatures, as well as the V-shaped opening at the back, were formulated by the machine to optimize the relationship between structure, weight and quantity.”
The true advantage of using algorithms to produce a chair, Luti explains, have to do with production. Even when you use traditional molds and not 3D printing. “First of all, we have used half a kilogram less of the material,” she specifies. “Which for a chair weighing about 5 kilos makes a significant difference. And we have also substantially reduced development timing for the product.”
Traditionally, from the design phase to the prototyping (also using rapid 3D prototyping) to the industrialization at least two years have to pass. “But not in this case. Although the AI development started in October, the chair was already in stores in July. Because the computer generates the optimal design to make the flow of liquid material perfectly cover the whole mold, distributing it to create the most functional solid structure. So the reduction of times, resources and materials is amazing.”
In spite of certain exaggerated statements by Starck (when Wired asked him “do you foresee the end of human designers?”, he replied: “of course, it is a prospect that does not scare me”), artificial intelligence does not represent a threat to human creativity.
“Quite the opposite,” says Anne Asensio, head of Design, Upstream Thinking, Innovation and Experience Design at Dassault Systèmes. Which at the latest FuoriSalone, inside the event “Design in the Age of Experience” at Superstudio Più, presented the Tamu chair designed by Patrick Jouin with 3DExperience (the above-mentioned firm that works with Airbus) using generative design.
“People have a phantasmagorical idea about machines. Actually, what generative design programs do is to create a space in which human beings can propose something new and improved with respect to what already exists. Within this scenario, the designer represents the intention, the invention. The software is a very powerful tool that lets him arrive where he could not arrive before. For example, to make an object have less ecological impact: to make it more compact for transport, lighter, using a smaller amount of material to achieve the same strength and durability. Generative algorithms also allow us to reproduce exactly the way a natural material grows and to imitate its becoming, in keeping with an approach known as biomimetics.”
This is precisely what happens with Tamu: the novelty of the chair lies in its replication of organic processes, making 3D design and printing easier while using a minimum quantity of material. And in the idea of a chair that folds back on itself, starting flat and becoming three-dimensional with a single movement.
So designers will not disappear. But they will have to evolve. “And they will have to change their way of thinking,” Asensio continues. “The challenge is no longer a matter of form and function, because with generative design we do not design only what can be seen, but also the structural nature of products: hence the material itself, which is invented based on the weight we want to obtain, the forces in play and the use that will be made of the product. For this reason, new designers will necessarily have to work in teams, with an in-depth understanding of synthetic materials, and some familiarity with biology.”
And they will develop a language shared with machines, which use a logic that is the diametrical opposite of theirs. Algorithms, in fact, proceed by analyzing data and operate within a well-defined box. “Designers, at least the good ones, think and invent outside the box,” Asensio says.
“Instead of clinging to positions in defiance of computers, they should try to understand what the algorithm does, to help them to be even more inventive. It is a question of background, and it is no coincidence that we are working with many schools all over Europe. All this will permit us to see the world with different eyes. The machines will be for design what photography has been for art: a glorious new start, a shift of perspective that has led great artists to explore beyond the visible, through Expressionism, Cubism, Conceptual Art.”
According to Daniele Speziani, CEO of Phitec Ingegneria (with a studio in Rivoli, known around the world and specializing in numeric simulations and generative design applied to product research and development, above all in the automotive sector), it is positive to talk about algorithms applied to furniture design. Speziani often intervenes ‘from the inside’ on products already on the market, calculating how to eliminate materials and redesigning parts that had been ‘over-designed.’
“There are errors that could be avoided. This is why it is positive to talk about these things: since this is a world unknown outside the niche of sector specialists, few designers and entrepreneurs know about the potential offered by generative design. It should be applied from the start of a project, to achieve results – in terms of optimization of resources and sustainability – that can definitely have a major impact.”