Frustrated with the misinformation swirling about hackers claiming responsbility for the removal of hundreds of sexuality & GLBT related titles from front page search results and the deletion of their sales rankings from individual book listing pages, and even more frustrated with Amazon’s lack of coherence on whether this was an internal error or an intentional omission of books from search listings, I spoke to an anomyous coder inside Amazon. That coder revealed that someone at Amazon — a real person — was responsible for tagging 58,000 titles as “adult” — not a hacker, and very likely not a “glitch,” either:
anoncoder: I can’t find the actual actual root cause, but it looks like someone internally changed 58K asins to be adult - whether that was accidental or intentional I couldn’t say, but we’re rolling it back.
So far, this makes sense — either someone, or someone’s account, is the origin of the mass application of the “adult” tag, which Amazon has already set to (wrongly, I editorialize) classify sexuality & queer books as “adult” and therefore to be excluded from front page searches.
Further debunking the hacker idea: one programmer tried out the troll’s instructions for how he allegedly hacked Amazon, and came to the conclusion that all this troll is exploiting is Amazon’s anti-moment. Another anonymous Amazon coder I contacted further explained the plausibility of such a hack.
another anoncoder: the way it was described in the livejournal link, yes, it’s possible. Tagging systems, whether Amazon’s or del.ico.us, are designed to make it easy to get lists of similar things. In this case, it sounds like someone used it to get a list of gay-interest items.
But that’s not an exploit, per se. That was an intended use for tags. And the actual code that the troll posted didn’t work. Part of that could be because Amazon plugged the holes, part could be that he was too lazy to post the entire exploit in working form.
And Amazon is still slow to roll out an official, or coherent, response. Feministing’s Jessica Valenti & her editor, Brooke Warner of Seal Press, got some answers out of an Amazon rep this afternoon. That explanation reverses the previous Amazon claim that all of this was just the fault of a “glitch”:
[The Amazon rep] also said no human is responsible for the decisions per se, and that it’s all about tagging and feeds which are constantly being tweaked. He does think that amazon will retweak the tags based on the uproar that happened over the weekend.
And in the time it took to transcribe all that, another internal Amazon source has come forward claiming that an Amazon France employee is the overzealous tagger. At this rate of blame-laying, maybe Amazon will propose by Tuesday that it’s all the fault of an as-yet-released Librarian AI, half-woman, half-machine. Which actually sounds pretty hot, right?
Update: Dear Author has started rounding up metadata on the affected titles: “Playboy Centerfold books were categorized as ‘nude’ and ‘erotic photography’” — both categories that apparently weren’t included in the filter” where Heather Has Two Mommies was labeled as GLBT. When GLBT was further filtered out as adult, this could explain why it was queer books that took the hit they did at the hands of whatever sexbot did this.
