Who Failed With #amazonfail?

I just read two very interesting articles from two commentators I respect immensely: Clay Shirky's The Failure of #amazonfail and Mary Hodder's Why Amazon Didn’t Just Have a Glitch. I won't do their arguments justice here, but I'll try to summarize as best I can.

Hodder argues that, even if, as it seems, the "glitch" is responsible for the recent #amazonfail twitter event, the algorithm itself contains values and assumptions, made by humans, that drive the results. I try to argue this to my students whenever this subject comes up: computers are not beings remote from us. They are us. Especially software, which represents an act of writing not very different from a polemic, a poem, or a magazine article--software, after all, is covered under copyright.

When a students says, "I'm not interested in computers, I'm interested in people" I always try to make it clear that this is a false duality. We make our computers--our "computing" is a cultural expression. One of my favorite projects over the past year was the work of my friend Christopher Martinez over at ASU. He talks about different, culturally sensitive ways of interacting with machines, a new form of computing based on practices that don't employ the typical keyboard-and-mouse Western industrialized mode. It's brilliant work and underscores how historically and culturally bounded our ideas of computers are.

Shirky argues that, even if this is true, the glitch is not the source of the #amazonfail outrage. He states that the scale of the reaction would not have been nearly as widespread or as virulent if a technical flub had been the source of the deranking of LGBTQ books. Rather, the anger comes from a sense of malice on the part of Amazon--hence the threat of boycotts and other actions at an entity held to be active persecuting its authors of gay literature.

Mmmmmmmmmmaybe. My objection is that the event got dubbed amazonfail and not amazonevil. Amazon failed to build a system that would preclude the pairing of "gay and lesbian issues" with "adult/sexual". Sarah Warn wrote a beautiful article during the "storm" that points out how often, in pop culture, the concept of "gay person" gets tied to the concept of "sexual", while the concept of "straight person" is not. My opinion, reflecting Hodder's, is that this equation made its way into the algorithm behind the glitch. I've worked with enough systems to know that this is a difficult thing for a computer to do by accident--someone with this idea built it in, even if unintentionally.

Shirky acknowledges that as a possibility, but finds it irrelevant. If the crowd behind #amazonfail had this in mind beforehand, that would've been one thing. Instead, they were reacting to something else. But, to take it back to #amazonfail versus #amazonevil, I think there was an acknowledgement, on the part of at least a large number of people, that Amazon's system failed. Yes, it failed because the people behind had a backward, if not outright evil, assumption, but the reaction was against failure, not outright persecution.

I'm sure I'm wrong about a pretty big portion of other people though--so maybe that's the point. We can't really say that the people contributing the Twitter during/on #amazonfail (is it an event? a body of writing?) were of one mind. I, for one, was initially interested in it as a phenomenon and later, when I decided that something bad had happened, reacted against what I thought was driving the algorithm by tagging up books. Yeah, puerile vandalism--I can't say I wasn't carried away by the spirit of the thing. But I do think, like the many other people who tagged deranked (and non-deranked) books on the site with "amazonfail" that we were addressing the (perceived) source of the injustice--a broken, manipulated, insecure algorithm that (we hoped) could be set right with group action.

Shirky is very right, though, that our hindsight is "20/400". So, don't take my word about my own behavior. Here are my posts related to #amazonfail. You can decide what my motivations were. And if you agree (or disagree) with me, post up your own results, and let's look at ourselves together, dispassionately, and see who's castle we thought we were storming.

Update: Fixed an embarrassing typo. mikefail.

Copyright Mike Edwards 2006-2009. All content available under the Creative Commons Attribution ShareAlike license, unless otherwise noted.