A significant portion of structural engineering tasks are lower-level tasks that AI and automation can easily complete. I'm surprised that it has taken this long. And I cannot wait.
Engineering design consists of two main phases: Discovery and Execution. Discovery is where the team works together to discover the structural solution to match the project success matrix, while Execution performs the engineering modelling analysis and documentation based on the discovered solution.
This Execution phase is where AI will take over.
This Discovery phase is where AI will not take over, at least not soon.
There are several standard architectural typologies that have matching structural solutions - what I call architectural-structural pairs (ArSt-2). The traditional two storey residential house for example, which is traditionally built with timber stick framing. These solutions don't necessarily require a Discovery phase for the structural engineering solution. The engineer can jump straight into the Execution phase. (I won't argue the point here, but that isn't necessarily the case.)
But where a standard pairing does not exist, a Discovery stage must be undertaken. And as engineers we are traditionally very poor at this. Often the solution is found by forcing an ArSt-2 when one doesn't really exist.
As the world changes and these ArSt-2 stop being "obvious", this Discovery phase is becoming more important. Mass timber is an example of the disruption that is occurring in these traditional pairings. And gone are the days when the structural solution for the "housing missing middle" was brick and concrete. This is an example of a lost ArSt-2. This has only occurred in the last 10 years or so and we haven't found a replacement. I argue that the traditional ArSt-2 pairings will eventually disappear altogether, including the beloved timber stick frame house. They just don't make sense any more. Co-creation will rule - where trusted ecosystems will develop solutions that are right for them, and not for the masses.
I am looking forward to AI taking over these lower level tasks, because this is mainly grunt work that generally doesn't bring much joy. The fun lives in the Discovery phase, were we get to play and use our skills to add value.
With AI, it will mean that it is even more important to make sure the tasks that we ask it to do are the right tasks. Because it will always spit out an answer. It will be important for us to be able to assess that we are asking it the right questions. Talented engineers making decisions that simplify structure will always trump complicated solutions that require complex structures that AI can deliver with ease. The fact that AI has delivered the solution with ease may mean we believe it is simple. But AI isn't going to be able to tell us that there is a simpler solution.
So lets focus our efforts on making sure we ask AI the right questions.
