Amazon Censorship

There are dangers to such large portions of our information infrastructure being in the hands of commercial companies in business to sell things.

Apparently Amazon’s “best sellers” lists are now in fact the best selling books that they (or more likely their poorly tuned computer learning algorithms) think won’t offend anyone. But perhaps more significantly, this apparently effects how books are ranked in search results too?

An LA Times blog writes:

Amazon’s policy of removing “adult” content from its rankings seems to be both new and unevenly implemented. On Saturday, self-published author Mark R. Probst noticed that his book had lost its ranking, and made inquiries. The response he got from Amazon’s customer service explained:

In consideration of our entire customer base, we exclude “adult” material from appearing in some searches and best seller lists. Since these lists are generated using sales ranks, adult materials must also be excluded from that feature.

Probst is the author of a novel for young adults with gay characters set in the old West; he was concerned that gay-friendly books were being unfairly targeted. Amazon has not responded to the LA Times request for clarification.

[Would it be better if the computer learning algorithms were more finely tuned?]

Some people are trying to establish the phrase Amazon Rank like so in Google, which as of this writing is working. I wonder if they should have chosen Amazon Sales Rank as the target phrase instead.


9 thoughts on “Amazon Censorship”

  1. I think some of has been happening for a while. I think you have to trigger some sort of thresholds or do certain things before being shown adult material. In particular I know I’ve looked for some comic books done by artists who I know have done some more…*cough* explicit material. Their mainstream comics come up, the adult ones don’t. There seems to be some complicated mechanisms here to block some of this stuff.

    Now the fact they all the sudden have classified gay romance and possibly other gay friendly books in this way however does seem to be new. I can’t help but notice that lately Amazon has been “lawyering up”. I imagine some right-winged group complained and so Amazon took the route that they thought would appease the right. Bad, bad decision on their part, but I do expect some lawyer figured that there’s more risk in a protest from the right-wing than the left-wing. It also could be a case of a rouge employee who is just going through those things and marking them as adult material.

    It’ll be interesting to see the shakeout. I’d hope they issue an apology on some of this and let those books show up on the bestseller lists and in searches.

  2. I don’t know if there are other reasons to cause this, but I find this particularly absurd when you consider affected non-fiction titles. From a very narrow bracket of topics, some examples:

    Philip Brett’s 1994 Queering the Pitch: The New Gay and Lesbian Musicology has no Sales Rank listed. Nor does The Queer Composition of America’s Sound: Gay Modernists, American Music, and National Identity by Nadine Hubbs, or Sedgwick’s Epistemology of the Closet.

    In contrast, Ruth Solie’s Musicology and Difference: Gender and Sexuality in Music Scholarship does have a Sales Rank. It makes me curious what the algorithms actually are.

  3. Yeah, the current algorithm (I’m still guessing this is automatically done, not humanly done) is particularly absurd. I suspect Jon’s right that this has been going on for a while, but perhaps has become even more absurd lately.

    It’s easier to argue against it (and certainly easier to NOTICE it) when it’s absurd like this.

    Would it be better (how much better?) if the algorithms were more ‘artificially intelligent’, less absurd, more consistent, more subtle?

    Liz, my guess is that the algorithms are NOT based on explicit rules, but instead are computer training type algorithms, such that you couldn’t really describe a logical decision tree or anything. That’s generally how automatic classification of any type is done these days.

  4. So maybe my question isn’t so hypothetical. It sounds to me like they’ve always had algorithmic censorship, but lately the automatic classifier got kind of absurd, leading more people to actually notice it.

    When they manage to return the automatic censor to it’s more subtle and less absurd state, more closely mimicing what a human censor would consciously choose — is there no longer any cause for concern?

  5. So, check this out. I don’t know if it’s true, but it’s QUITE plausible.

    Rather than automatic classification, simply based on user’s votes for what is ‘inappropriate’? Even assuming this was somehow protected from ‘hacking’, I’m not sure I like the idea of Amazon censoring books based on votes either.

    If that’s really the way the system works, Amazon is in for some interesting times. Don’t like a book? Get all your friends to vote it ‘inappropriate’, it’s made less visible on Amazon, hurting it’s sales. No need for actual cross-site hacking techniques if you’ve got enough friends or comrades who don’t like it and want to hurt the author.

    This system is problematic. If that’s how it works, it was problematic when it was just happening to a few freaky authors here and there too. If that guy did what he says, I say he performed a service by bringing this issue to attention.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s