Artificial intelligence

Artificial intelligence 

I do expect a lot from artificial intelligence. First of all, I'm sure that low-level practical intelligence can liberate human designers from trivial or repetitive tasks, like the proper positioning of a door, so that they can concentrate on the bigger picture, e.g. how circulation takes place in the design. One should obviously know all about doors and their positioning but shouldn't have to spend too much time on each door in a design.

In addition to this small yet essential stuff (and rather tricky in terms of intelligence), I want even more from computers, especially feedback from analyses that monitor design actions and decisions, calculating their impact on behaviour and performance, so as to give early warning to the human designers. The analyses and simulations already exist to a large degree but have yet to be integrated in designing as they should be for this kind of immediate and continuous feedback. Analysing a design once it's practically finished contributes little to improving the design.

If this partnership between the analytical computer as an intelligent design assistant and the creative human designer seems obvious, one can't say the same about intelligence that allows computers to understand design intentions through types, grids, schemata etc. and impose these as coordinating devices. It may seem strange that an analytical assistant should take over the design like that but there's a lot an assistant can do to promote consistency, including assuming such a coaching role. This can also prove liberating to designers, letting them focus on intermediate levels (the spatial aspects that often go wrong in buildings), as well as making them see the overall picture that emerges from their possibly uncoordinated actions. Who know what kind of clever distortions and antitheses they might conjure up once they realise it.