Less Than 20% Of People Building AI Are Women. Here's Why It's A Problem.

I worry a lot about AI, but not in the ways that most people worry about AI — the fourth industrial revolution is already in progress, but we're still light years away from that unknowable moment where we create something that truly thinks for itself. Where we are, now, is creating machines that are better (and, sometimes, worse) at learning how to do human-like activities. Algorithms are an everyday part of our lives, as are bots; to the extent that we're probably not even aware of them half the time (Google, anyone?). This is the level of so-called 'AI' most of us will need to confront in the near future. And this is where our learnings, now, will inform what might come to pass in the not-so-near future ...

One of the biggest concerns with algos and machine learning is how, as these functions are all human coded or designed, they reflect the human biases of their programmers or coders — except that they can now do this much faster. Many of the tech stories I've recommended on this platform have looked at race bias in everything from facial recognition programmes to code that tries to predict criminal recidivism rates (both do it badly, FYI).

I've been overlooking another really key issue, and that is: the massive lack of women in tech overall, and how this effects, for example, even the subservient bots we currently have in place — and how this will no doubt impact on the higher-thinking machines and programmes we will have in the future.

This article (which does read a bit press releasey at times) is an excellent summary of just the very basic things that are overlooked, including female anatomy, and the inherent (and scary) sexism that is tacitly imposed on algos almost certainly coded by almost all male teams.

It's not a future I want to live with, which means we need to think seriously about how to change the ratio.