Skip to main content

Amazon's Harm

As a software engineer, and a leader of software engineers, I often think about the user experience. Almost every piece of software I have worked on, or been responsible for, has had a connection to the average consumer. Awareness of the consumer is critical.

A trans author, Shon Faye, has written a book called The Transgender Issue: An Argument for Justice. It is extremely well regarded, addresses a number of issues facing trans people, and I want to read it. Shon is born, and resides, in the UK, so the book has been published there. In any event, lots of books written for and against trans people have come from the UK and I figured I would be able to buy it here...

Well, turns out that is not only not true, but when I search Shon Faye on Amazon, they return to me a list of recommended hate screeds by bigoted UK transphobes instead.

Think about that for a second: Amazon knows who Shon Faye is and that she is writing trans material. 

Amazon has built their data and query infrastructure to presume that people searching on trans issues are materially looking for anti-trans writing and they optimized results for that. So, instead of showing me results from other trans authors, they showed me bigots.

More specifically, they showed me bigots who are explcitely seeking to harm and exclude trans people from society. They write to generate hate and disgust of trans people and they don't really care if trans people are hurt in the process. Somehow Amazon sees fit to carry that in Canada, but not the positive trans writings of a trans author. Then they flog that material to people looking for trans positive writings. I call that harm.

It's definitely harm. Good luck getting through to anyone at Amazon to actually address it though. Have you tried their "Contact Us" options? LOL, right. This is exactly how harm perpetuates.

So what this really comes down to, as it always does, is that the vast majority of this software is written by cishet white men and they think of none but themselves. The evidence of that is all around us in machine learning, facial recognition, and search algorithms. 

It's continuing to cause harm.


Comments

While you're here, you might like:

I identify as... Stop! It's what I am.

The expression "identifies as" gets used a lot when talking about trans and non-binary people. It's also used by a lot of bigots to denigrate the trans experience. It's an expression that I would like to see fade away.

When a company (Voilá) keeps screwing up

Now it's time to really talk about this because fumble after fumble has been the name of the game for Voilá by Sobeys. This story is the basis of my earlier post:  It's not polite when you get it wrong .

It's not polite when you get it wrong

I've talked about voice issues before, for many of us in the trans community it remains a major source of frustration and irritation. This is especially true on the phone.