What is the AIAD revolution?

What is the AIAD revolution?

Par Samuel Weiss

The creative process and approach to a new project varies from architect to architect. Some deal first with the history of the site and society, others look for references from projects with similar typological requirements, others use regional climate data and still others first look at the infrastructure surrounding the project to determine their design parameters. The final language of architecture, aesthetic norms and avant-garde forms, however, have always been created depending on the technical possibilities an architect has to capture his creativity and ideas. Everyone now knows that our world is changing rapidly. Our everyday lives have already been revolutionized by big data and artificial intelligence, and many aspects of society are changing. For example, the future of our world of work depends on how we manage to use AIs as cognitive partners – so for us as designers, too, it‘s about finding ways to integrate artificial intelligences to enrich the architect‘s design routine.

What can contemporary digital design tools do – and what can they not?

Until recently, designers only had passive tools available for their work. They did exactly what they were told to do and nothing more. The introduction of algorithms into everyday design life did nothing to change this – only the variability of the designs was extended: instead of a single object, the design process could now be designed with all its dependencies and thus generate a large number of objects in parallel. Today, intelligent design programs are (still) limited to Big Data and simulation. In this way, shapes can be optimized or generated in relation to certain parameters in a very uncomplicated way. Generative design tools even go one step further: They only need the fixed initial parameters such as goals and constraints to generate a multitude of possible variants. Human beings are still responsible for the final evaluation of the results – according to purely visual criteria.

How can an AIAD programme enrich the architect‘s working process?

The creative action of artificial intelligence currently consists only of creating artifacts such as songs, paintings or film scripts. The result, however, always depends on the data set used as input. While they are able to understand and reproduce complex patterns better than humans, they do not understand what these patterns mean. If creative collaboration between people and machines is to be encouraged and enriched, the best of both ways of thinking and working must be united. The consistent integration of artificial intelligence in CAD programs and thus in the creative process of the architect elevates an AIAD program from a pure ‚tool “to an employee” or “personal assistant” who makes suggestions and gives design relevant advice. With a collaboration of this kind, it would be possible for a new language of form to actually emerge in architecture – and not just a machine-calculated increase in material efficiency. So the question is: does architecture adapt to technology or does technology adapt to architecture?

This phase of collaboration between man and machine serves as a learning phase of the AI towards a completely autonomous work of AIAD programs. Such a scenario makes the architect as designer obsolete and at the same time raises the question of who is responsible for input and design decisions. For a perfect ‚User Generated Design” only the interface of the programs has to be designed user friendly – the designers” of the future are no longer architects. Here I will show two different ways: On the one hand, thanks to the processing of huge amounts of data, it will be possible to let the affected population group decide on architecture in a grassroots democratic way. On the other hand, the design-relevant parameters could be staked out and defined using a kind of questionnaire within an app developed for this purpose, or the personal accounts of Instagram or Pinterest could be used to generate a wide variety of designs from which the user can then choose his favourite.

Lire le mémoire complet :