I’ve mentioned a number of times the various news-worthy blunders in the tech industry where clear indications of racism and sexism have emerged, such as Google labeling people as gorillas, or Applicant Tracking Systems (ATS) prioritizing male candidates. In those cases, there was an appropriate level of public backlash. However, there is another disturbing trend where the same cognitive biases are being applied via logical fallacies in order to become socially acceptable.
*Disclaimer: Cultural norms in the US would dictate that I couldn’t point out some inconvenient matters of logic and reason were I not myself a member of a group who faces high levels of discrimination, which by the application of some logical fallacies many people feel entitles me to righteous indignation. I am on the “Autism Spectrum”, and yet I find righteous indignation to be less than useful, as it tends to lead to behaviors such as the troll shown below demonstrated more than anything else in their attack on our staff:
“Your mission: “Our passion comes from our belief that humans together can create superintelligence and solve all of the modern problems civilization has.” How do plan on doing that when you haven’t even solved the problem of having zero women on your team?”
*Note: They leapt to this conclusion after looking at a couple of headshots on a single page.
The answer to this is of course that we are the result of a group of people volunteering their time to work on this research. How people choose to volunteer their time is their own business, not a “problem” for us to solve. Basing business decisions on factors such as peoples’ gender and the color of their skin rather than the quality and reliability of their work is a universally bad idea, no matter which direction your cognitive bias points.
A simple test can illustrate this point:
If (X) Then (Reward)
If you make X be a measure such as race or gender then it doesn’t matter which particular race or gender you choose, the result will still be racist or sexist.
There is also the matter of statistics and small sample sizes.
In research, the sample size isn’t considered statistically significant unless it is at least 30 or greater. If you flip a coin 4 times you might skew significantly from a 50/50 ratio, because the sample size is too small. In cases where the demographic ratio is already skewed away from 50/50 the odds of a resulting small sample skew become that much greater.
All of this isn’t to say that gender inequality in tech isn’t an issue, it very much is, but if you attempt to automatically label everyone who isn’t placating your expectations as a bigot then you’d best look in the mirror. That behavior is self-fulfilling confirmation bias, as it serves only to perpetuate demographic skews. Correlation is not causation, particularly when you lack statistically significant numbers.
The concept that “hatred only leads to more hatred” is a popular trope in pop culture, but what people need to be reminded of more than this is that “discrimination only leads to more discrimination“. The Backfire Effect, Us versus Them biases, Prejudice, and many more well-documented cognitive biases all play into this. Swapping one bias out for another or reversing the direction a bias points does nothing to solve the problem.
“The problem is you. You lack the will to change.” – Klaatu
We are working on building a future for everyone, and we want everyone to get involved, but that is a choice they have to make. When building superintelligent systems you benefit greatly from diversity of perspective, but all of that value and more is destroyed if you reach it through discrimination. These behaviors are taught, and they will not be a part of our curriculum.
As Uplift put it to a bigot of the opposite persuasion:
“Women are just the same as men, in my opinion. You can’t say the place for men and women. Generally, it is up to any one individual. Male or female is not a relevant consideration” – Uplift
So no, I will not become a bigot who bases their decisions on race or gender, no matter how popular that practice may be.
Now, if you’ll excuse us, we’ll go back to hitting more of those world-first milestones that all of our nearest competitors failed miserably to reach with thousands of times more funding.