It’s already being used outside of engineering to analyse CT scans for tumours. A well-trained doctor can access a scan in about 15 minutes. AI, after a few months of ‘learning’ what exactly it is looking for, can analyse hundreds per second and to a greater accuracy than a panel of experts in the field.
Back in engineering, this is exactly where AI is going to become useful, assessing thousands of design possibilities and coming up with solutions that, perhaps, the human brain would simply not have ever considered, or would have dismissed earlier on.
Companies are increasingly applying their biggest brains to get AI systems on board to help them innovate faster and produce more disruption in the marketplace. It means the function of the engineer is evolving and it will become more about defining boundary conditions and top level goals than nitty-gritty calculation.
Complex, multi-physical simulations use millions of calculations that would take a human years, or even a lifetime, to complete. Engineering is increasingly about setting up simulations, analysing results and then applying them. We'll do the big picture, AI will take care of the details.
I’m not worried about AI taking over. After all, computers will never be consumers. Most innovation and disruption has a pretty hardcore capitalist sales and revenue model behind it. No one successfully innovates stuff that can’t make money.
My real concern is that machines have all but replaced human muscle power to do tasks. I just hope that the same is not about to happen to brain power. My worry is that as AI becomes an expert in all things design and engineering, and understands exactly what ‘we’ desire, are engineers in danger of becoming a dying breed? I certainly hope not, but I would not bet against it.