TNW Conference 2022 will be bigger, bolder, and better! Get your tickets now >>

Human-centric AI news and analysis

This article was published on September 7, 2021

Facebook is very sorry that we keep noticing its racist AI

Making AI that isn't racist is too hard and not using racist AI is apparently out of the question

Facebook is very sorry that we keep noticing its racist AI
Tristan Greene
Story by

Tristan Greene

Editor, Neural by TNW

Tristan covers human-centric artificial intelligence advances, politics, queer stuff, cannabis, and gaming. Pronouns: He/him Tristan covers human-centric artificial intelligence advances, politics, queer stuff, cannabis, and gaming. Pronouns: He/him

Did you know Neural is taking the stage this fall? Together with an amazing line-up of experts, we will explore the future of AI during TNW Conference 2021. Secure your ticket now!

Facebook’s issuing apologies for its AI‘s bigotry again. This time, users watching a video of a Black man were asked it they were interested in more content on “primates.”

The company, by way of apology, said:

As we have said, while we have made improvements to our AI, we know it’s not perfect, and we have more progress to make. We apologize to anyone who may have seen these offensive recommendations.

Up front: This is nowhere near the first time Facebook’s had a problem with its AI‘s bigotry.

Whether it’s hiring algorithms that are flat out misogynistic, translation errors that result in wrongful imprisonment, or its blanket discrimination against queer people, Facebook’s AI is broken for everyone except straight white men.

Background: In fact, the only thing surprising about this particular problem is that it happened to Google about six years ago.

And that’s the exact lens we should view these issues through when Facebook promises its algorithms will improve. When? At what point is it unacceptable to have product that, at its core, is discriminatory?

Realistically, Facebook’s apologies and promises are less than worthless: they’re harmful.

You wouldn’t sell a garage door opener that didn’t work for Asians. You wouldn’t install an elevator in your building with a sign that said “only works for white men.” And you wouldn’t purchase a driverless car that struggled to identify Black people as pedestrians.

But Facebook would. It uses myriad AI systems that, time and time again, demonstrably struggle to deal with women, Black faces, queer people, and almost any other marginalized group.

Quick take: Facebook’s valued at over a trillion dollars as of 2021. If it chose to only launch products that weren’t racist, it wouldn’t be.

Facebook willfully releases products that are inherently discriminatory because it feels that profiting off of racism is more acceptable than not profiting. Building algorithms that aren’t discriminatory is, apparently, beyond the scope of modern technology.

As long as companies and their leadership teams are willing to peddle racism for profit, racism will remain profitable.

[Related: Facebook apologizes. It does other stuff too, but mostly it just apologizes]

Get the Neural newsletter

Greetings Humanoids! Did you know we have a newsletter all about AI? You can subscribe to it right here.

Also tagged with