Its a combination of two separate problems- the skewed demographics of software engineers and the continued requirement of an intuitive understanding of the pro

Its a combination of two separate problems- the skewed demographics of software engineers and the continued requirement of an intuitive understanding of the problem AI is being developed to solve. If you're building an AI system to help police predict crimes, you want to have a detailed enough understanding of the places its going to be applied to know what the impact of its predictions would be and how to shape the system to actually produce a desired result. Like with UX design, going from a specification to a logically valid implementation isn't necessarily going to produce results people actually want. Its much the same way as, say, startups trying to market products to new mothers or highschool students have a hard time because they can't "eat their own dogfood" if they're all late-twenty-something white men. Huge markets lie unserved, I'm sure, because they're demographics that don't tend to produce software engineers or entrepreneurs.