Saturday, November 7, 2020

Movie with Abe: Coded Bias

Coded Bias
Directed by Shalini Kantayya
Released November 11, 2020 (Theaters and Virtual Cinema)

So much of life in the present moment is spent on devices. People have their phones with them at all times and consume tremendous amounts of media on their computers and Smart TVs. Advertising is very prominent on all platforms, and it’s not uncommon for Facebook or other services to draw from Internet searches or conversations to make targeted recommendations that can make even the most pro-technology person feel like they’re being watched. But there’s also something far more disturbing embedded within all these algorithms that is being used to expand systems that, based on the information that is being learned, will only serve to perpetuate the inequality that currently exists.

Joy Buolamwini is a researcher at MIT Media Lab who made an important discovery: as a Black woman, she was not being identified by a facial recognition system. When she put on a white mask, it immediately found her. Further studies show again and again that there are inherent biases that are inherently part of artificial intelligence systems. They learn information from what is already out there in the world, and the way that they digest and use that data is often purely mathematical, meaning that what is already true will simply be replicated. Left unchecked, there are disastrous implications, and, even more disturbingly, many large and powerful companies have no desire to do anything to counteract them.

This film is a fascinating – and worrisome – study of concepts like machine neutrality, meant to mirror objectivity but instead serving as yet another propagator of the bias that is already at play related to gender, race, and many other factors. It also surveys algorithms for teachers or parolees that are not able to factor in positive elements like accolades and recommendations and therefore set someone up for failure by ensuring that they are held to the worst standards, regardless of their accomplishments. There are many applications of this, and it’s clear that this is just skimming the surface, with far more dangerous uses to be found in the near future.

This film works well because it doesn’t pretend that its audience is either too intelligent or not intelligent enough, explaining how the algorithms work without going into technical detail that might make it more confusing and confound the issue. It serves as a strong and important call to turn back before it’s too late, emphasizing the need to protect the right to not be captured on camera so that the United States can start using its artificial intelligence systems for political intimidation purposes, as has been displayed in China. There’s a great deal to be learned here, and this film is an excellent and vital introduction to something that is far more advanced and problematic than an unconcerned public seems to think.

B+

No comments: