Adults Lose Skills to AI. Children Never Build Them. (Psychology Today, 3/22)

[Johnny2X2X]

55. This is going to be a big problem

In my line of work, Aerospace, specifically avionics, young engineers learn by doing. And for several of our products, learning means testing. We have junior engineers write test scripts for the hardware, software, and systems we sell. It's one of the first things you'll be assigned to do when you start, you'll get one little set of requirements that describe some function for you to dive into and figure out how to verify. Might take weeks or months of learning about the desired behavior and what our system does, but at the end of it, you'll be a little bit of an expert on that small part of our product and you'll move onto the next set. After a year or two you'll have a pretty deep understanding of both our requirements and how they are implemented in our systems. You'll learn about process and coding standards, you'll learn about robustness testing and for credit test execution. You'll schedule time with senior engineers who are the experts and get chances to pick their brains.

One of the first things we're doing with AI is trying to automate writing test scripts. Now process for certifying airplane parts dictates a human will have to validate (through peer review usually) test scripts, test cases, and test results, but the junior engineer usually sits as author on those reviews. We're cognizant of the loss in development here and are working plans to get young engineers other experiences, but it can't just be a hand waving training exercise, these junior engineers need hands on experience solving real problems for themselves or they won't become tomorrow's experts that innovate and push our technology forward.

Companies who can best utilize AI while still cultivating human technical expertise will be at a competitive advantage. Who can figure out how to walk that line best will win in their market.