Machine learning is the new fidget spinner in IT circles, it would seem. The only problem is, those darn machines are sexist, racist bigots, just like the rest of us.
The article above explains how systems such as Google’s Sentiment Analyzer are producing results that infer a negative bias against certain groups based on ethnic, sexual or gender identifying nouns.
Everyone seems quite surprised and somewhat disappointed by this discovery.
Perhaps what should strike us as most strange about this is that anyone would predict that these systems would be unbiased.
Thinking about the root source of the learning material of the algorithms; all they have to start with is human speech and the written word. The programmers have let the software loose on the collected wisdom of mankind and asked it to draw its own conclusions.
Unsurprisingly, the software has discovered that we all use bias and we all use it all the time.
Perhaps the next conclusion the algorithms might offer is that bias is an entirely natural, logical and, indeed the only known way for humans to successful navigate the world.
“Bias” is a synonym for “in-group preference“, that is, the system every single one of our ancestors employed to stay alive.
Fear or careful suspicion of animals and plants of unknown species would have kept your and my ancestors alive on the plains of Africa long enough to mate and have offspring. That same fear and suspicion of other humans outside of their immediate group also protected our ancestors from being “victim zero” in the next inter-tribal raid.
Later in our evolutionary history, communicating at a distance with those outside of their immediate group will have saved countless of our ancestors from deadly diseases against which their genes hadn’t yet developed an immunity, again, allowing them to mate and have offspring.
How do we know this was a highly-successful strategy that beat all other competing strategies attempted by their peers?
Because I’m here and able to write this blog post and you are able to read it.
There is nothing shameful about bias, per se. It has served us well throughout every previous generation. Irrational bias is, by its nature, illogical, but before we write off every momentary expression of in-group preference as racist/sexist/whatever the current “-ist” du jour is, we might consider whether it is actually irrational or whether there is any utility to be had by employing it.
In the words of G. K. Chesterton,
In the matter of reforming things, as distinct from deforming them, there is one plain and simple principle; a principle which will probably be called a paradox. There exists in such a case a certain institution or law; let us say, for the sake of simplicity, a fence or gate erected across a road. The more modern type of reformer goes gaily up to it and says, “I don’t see the use of this; let us clear it away.” To which the more intelligent type of reformer will do well to answer: “If you don’t see the use of it, I certainly won’t let you clear it away. Go away and think. Then, when you can come back and tell me that you do see the use of it, I may allow you to destroy it.