A lot of what we call “engineering judgement” might just be pattern recognition.
The experienced engineer who walks onto a site and immediately senses something is wrong. The one who looks at a drawing and knows before running the numbers that the scheme won’t work. That intuition is incredibly valuable, but is it fundamentally different from what a model trained on enough similar situations could develop?
I want to say yes. There are elements of reasoning, ethics, communication and contextual understanding that aren’t easily captured in pattern matching. But I’m honestly not certain, and I am massively biased as someone who has built up 10+ years of this intuition. I’d be suspicious of someone who’s very confident either way.
What I do think is that if judgement is substantially pattern recognition, then AI has a more credible path into the core of this profession than a lot of people are comfortable admitting.
The big but is that it needs enough data and enough time.