This a Guest Post by Mary Hodder, founder of Dabble.com, a social video search site and blogger for Napsterization. Hodder is a veteran Silicon Valley technologist and was most recently VP of Products at Apisiphere, a geolocation mobile company building an enterprise platform for mobile developers. This post is in response to the Amazon’s removal of sales rankings on a number of gay- and lesbian-themed books due to a glitch.
Webopedia defines an algorithm as:
(al´g&-rith-&m) (n.) A formula or set of steps for solving a particular problem. To be an algorithm, a set of rules must be unambiguous and have a clear stopping point. Algorithms can be expressed in any language, from natural languages like English or French to programming languages like FORTRAN.
We use algorithms every day. For example, a recipe for baking a cake is an algorithm. Most programs, with the exception of some artificial intelligence applications, consist of algorithms. Inventing elegant algorithms — algorithms that are simple and require the fewest steps possible — is one of the principal challenges in programming.
The ethical issue with algorithms and information systems generally is that they make choices about what information to use, or display or hide, and this makes them very powerful. These choices are never made in a vacuum and reflect both the conscious and subconscious assumptions and ideas of their creators.
The ethics bar in creating algorithms and classification systems should be very high. In fact I would suggest that companies with power in the marketplace, both commercial and ideas, should consider outside review of these systems’ assumptions and points of view so the results are fair.
Algorithms are often invisible, and difficult to detect by design, because technologies that use them are designed not to share the methods for providing information. This is mainly because users are focused on the tasks at hand in information systems, and given good information, they don’t need to know everything under the system’s hood, and because technology makers like to keep the ’secret sauce” hidden from competitors, not to mention people who would game systems for their own devices such as spammers or other bad actors.
However, the flip side of this is the lack of notice to users of the assumptions that a system is making, which may not even be apparent to those building the system. Systems that filter out useless, spamming or other uninteresting information also can filter out other things without notice. When a page is full of data, it’s often very difficult for someone to recognize what is missing. This is a sort of “where’s Waldo” situation except that in the mass of data shown, there isn’t a Waldo. Waldo is invisible.
This weekend it was discovered that Amazon had filtered out some books as “adult” even though those books had minor sexual or erotic themes (fiction) or were non-fiction discussion classes of people associated with particular sexual orientations. Apparently, these filters began to pop up two months ago and while Amazon was notified at the time, many — including me — weren’t aware of the problem until Mark Probst, author of “The Filly,” realized his book was missing it’s “sales rank” from Amazon. He questioned them, and they replied:
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.
Hence, if you have further questions, kindly write back to us.
Best regards,
Ashlyn D
Member Services
Amazon.com Advantage
Sunday afternoon, I noticed after investigating the issue, that twitter had about 500 posts every 15 minutes with the tag: #AmazonFail. A petition appeared with 400 signatures and quickly jumped to 2,000 within a couple of hours, and then to 10,000 by Monday morning. Tweets were at a high of 6,500+ midday Monday over a 45 minute period.
Using the tagging system in Amazon, users tagged about 1,000 books that generally have gay and lesbian themes with AmazonFail (whether non-fiction historical or fictional stories, whether soft-erotic — think Jackie Collins level stuff — or documentary on things like the military policy “don’t ask don’t tell”). Books like Lady Chatterley’s Lover, and The Mayor of Castro Street and Maurice by EM Forester and Conduct Unbecoming, a history of gays and lesbians in the military, were included.
Here is a comparison example for the details section of Amazon pages, where one book has it’s sales rank by Amazon, and another is missing:
Note that the second book, A Parent’s Guide to Preventing Homosexuality (Paperback), has many reviews (most of which have gone up in the past 48 hours and are negative) in part I think because this is the top result for the search: “Homosexuality” in Amazon’s system.
Note that Amazon came out late Monday with the explanation that this was all “a glitch” in a statement to Publisher’s Weekly, however that contradicts earlier email from them to authors stating that they were in the “adult” category simply for including positive gay and lesbian themes in their works and that’s why they lost their “Sales Rank” statistic that would keep them in search results. It was a very targeted glitch for sure. Targeted to, among other things, “positive references to sexual orientation == gay” placing them into the “adult” category, which allowed the other minor “glitch” by the programmer to be possible.
If all this seems like a problem, and it should, it’s because Amazon is using algorithms, which rely on their classification system, with various statistics like “Sales Rank” to rank products in search results on the site. These algorithms and classifications have points of view. Their point of view, revealed this week, is that “positive references to sexual orientation == gay” is “adult” in nature. And that classifications will be used in the algorithms to sort out what is shown and what does or does not get to have “sales rank,” which then orders items in search results. And we all know search result order can lead to big sales, or invisibility. The SEO industry and Google bank big on that point.
Search for “homosexuality” in Amazon and this is the top result:
While Amazon “Sales Rank” numbers in part determine the ordering of search results, I’m guessing that the number of user activities around a listing, such as reviews, may also be a factor because yesterday when I did this search, the 5 books ranked following “A Parents Guide…” were also anti-homosexuality. Today, the four that follow are positive toward homosexuality (at a glance), and from what I can tell there are many new reviews and comments on these books.
The issue with #AmazonFail isn’t that a French Employee pressed the wrong button or could affect the system by changing “false” to “true” in filtering certain “adult” classified items, it’s that Amazon’s system has assumptions such as: sexual orientation is part of “adult”. And “gay” is part of “adult.” In other words, #AmazonFail is about the subconscious assumptions of people built into algorithms and classification that contain discriminatory ideas. When other employees use the system, whether they themselves agree with the underlying assumptions of the algorithms and classification system, or even realize the system has these point’s of view built in, they can put those assumptions into force, as the Amazon France Employee apparently did according to Amazon.
This of course doesn’t explain how the problem arose two months ago, and why when Amazon was notified, they didn’t look into it then. I would suggest that the same underlying assumptions that drove their classification and algorithm system to be built to filter “gay” into “adult” also led their investigations in February and March to lead to nothing. It was only public outrage this past weekend that caused them to look harder, beyond their own assumptions, to find the underlying problem.
The bar for ethics in creating algorithms and classification systems should be very high. #AmazonFail proved it’s not, at least at Amazon. I would venture that Amazon’s classification and algorithm system have more of these discriminatory assumptions, and while their tagging system does allow users to correct for some of this, Amazon is using it’s internal classification system in it’s filters, not user tagging, that I can tell.
I would suggest that the company, because of its position in the market and power over both authors and publishers, as well as users and the intellectual marketplace of idea, ought to be doing a complete and public review of their classification and algorithm assumptions. Publishers and authors should push for it, and so should users.
For details on the amazon exploit by the guy who did it, see the pastebin here: http://pastebin.ca/1390576
@Chris,
That “exploit” was debunked. Just a troll looking for attention.
Really? Where?
Yeah, it was some trolls from 4chan. It’s amazing what you can do on the Internet with a few iframes and a CSRF hole.
I agree that “gay” does not always imply “adult”, but “gay” implies “adult” about 95% of the times.
If you filter “gay”, you only have 5% of false positives. which is great. If you keep “gay”, the system becomes flooded with adult stuff.
Algorithms work on data. Algorithms are not wrong, they just reflect what they reality is.
That’s probably what the author of this post is struggling with: accepting reality.
If “gay” is adult 95% of the time, than Ron Jeremy is adult 99% of the time, yet his biography was not affected by the “glitch.” How is “Heather Has Two Mommies”, which was de-ranked more “adult” than Ron Jeremy’s bio or a photographic book on Playboy centerfolds which both remained ranked?
The fact is, the majority of the books de-ranked were LGBT oriented books while cis, hetero books mostly went untouched.
I dunno what your agenda is but I aint buying it.
Actually “gay” does not mean “adult” 95% of the time and keeping gay themed books in the search does not mean that Amazon will be flooded with adult material. You may be an engineer but your view of reality here is extremely skewed.
It’s not worth mentioning that an “algorithm” is in play. Actually it’s condescending to define it here, where it’s techcrunch, and the tenth grade vocab list. Sorry, I’m biased against snobs.
I have spent thousands every year for the last ten years at Amazon and touted Bezos as the tops the whole time. I don’t believe a whit of what they’ve said and, regardless of what they do, I will never buy from them again. I can’t wait to see if Barnes and Noble, or anyone, hires someone to go after my business.
Well said Mary – it’s so great that this mess was exposed so fast.
Just a few years ago it was so easy for a company with Amazon’s power to influence society to push one ideology or another and sweep their bias under the rug – with nobody even noticing.
A couple lines of code, and boom – only “anti-gay” books show up. Another quick tweak to the code, and only “abstinence only” books show up, to heck with sex education, because maybe some religious fundamentalist programmer/analyst somewhere don’t think it should be taught – so there, it’s gone.
This type of ideology based censorship is very dangerous, and the faster it’s exposed, the better. Yeah, so right – Amazon needs to do a very public review of their internal classification system assumptions – because until they do, lots and lots of people are going to assume this was Amazon’s stance and they backpedaled when it hit the fan.
There’s a quote I love: “sunlight is the best disinfectant”. How true.
Have an awesome day!
Dan (Ask Dan and Jennifer)
Dan & Jennifer said…
This type of ideology based censorship is very dangerous, and the faster it’s exposed, the better.
Dan, here is a short answer to your misguided censorship that you’ve just posted above, which I pinched from this blog post.
Quote:
—–
Censorship is interference by the state in the expression of ideas. It is not censorship if a free agent, eg., a newspaper, decides to edit or reject your copy. There is nothing in the principle of free speech requiring that I provide you with a megaphone, or the Herald newspaper provide you with a platform. A private network refusing to publish your views is not censorship – it is their choice. A private newspaper editing your ill-thought maunderings is not censorship, it is good judgement.
So, what is Amazon (or actually any private enterprise such as Google and others) is doing with what they own (ie, legitimately owned property) is their choice not censorship. So, I hope you understand the difference.
Umm…censorship can be done by any institution, organization or the government. The only difference is that the government is “banned” from practicing censorship, and private companies and organizations are not. If any company, even Amazon, wanted to practice any form of censorship, they are free to do so, but they must accept the consequences of that action, good or bad.
LOL, let’s clarify. Indeed – if Amazon (or whatever other private company) wants to hold a press conference tomorrow saying that certain books – say gay-themed books or books about people of a certain color, or about sex education – are not welcome on their site anymore, that’s absolutely their prerogative and their right on their site, though it could be an expensive decision, and one they’d probably never make.
The point here is a fine one, it’s ONLY about exposing subtle “behind the scenes” type of censorship that’s done on the down-low, in a way that people probably won’t notice, whether intentional party-line or “accidental”. Again… bringing it out into the open.
Have an awesome day!
Dan (Ask Dan and Jennifer)
Here’s another example – “taking away one small freedom at a time”.
You don’t take all their freedoms away at once, they’ll rebel. But if you only take away one freedom at a time, and give them a good scare / reason for each, then all you need is patience – just hope nobody notices the plan.
That’s the point to visibility… keeping the little things in the light before they blossom into bigger things.
– Dan
DougF said…
If any company, even Amazon, wanted to practice any form of censorship, they are free to do so, but they must accept the consequences of that action, good or bad.
Exactly, that’s what you call free-market. The market decides what a business can or can’t do. Wrong decisions and you fall and when you make good decisions then you rise, simple as that. There is no need for government interference at all. The market is the regulator itself.
Amazon is in the business of selling books : from mass murder, homosexual, science fictions, science , non-fictions , etc… There is no such thing as deliberate exclusions of some titles, since they’re the one who’s losing the sale. Their recommendation engine tries to up-sell & cross-sell to a wider varieties of customers and they do know that there are kids or parents searching for kids title. They probably try to make their recommender to filter out those parents/children’s searches by excluding titles that include : homosexual, adult, gay, etc,… Excluding these titles from children/parents search will likely to keep them searching the sites (potential for a sale). If the recommender spews out titles which are offensive to parents/children, then the likelihood of continue searching with the site drops.
Marry Hodder said…
I would suggest that the company, because of its position in the market and power over both authors and publishers, as well as users and the intellectual marketplace of idea, ought to be doing a complete and public review of their classification and algorithm assumptions.
That’s gonna be a very difficult ask Marry. It is like asking Amazon to tell us (the public) what their algorithm is. A similar case between Google and SearchKing a few years ago in which Google won, SearchKing lawyers were asking the judge at the time that they need to examine Google’s PageRank algorithm to find out why Google search engine de-ranked SearchKing’s affiliated sites.
Marry Hodder said…
Publishers and authors should push for it, and so should users.
It should be done on a voluntary basis since Amazon is a private company. Lobbying for legislation to force Amazon to do this is not the solution.
This seems like a bunch of bellyaching. Amzn is leading the e commerce revolution with great deals, and its stock keeps surging. I think they will be fine. http://iamned.com/blog/
And the fact that their stock is surging serves as ultimate proof that whatever they do is moral and ethical? Hm.
thank you for pointing that out. amazing how shallow our world has become. i am shorting AMZN and if anyone doesn’t think this goes deep, they’ve not see how much bigotry is hated.
No, but it is reassuring to know that a bunch of reactive whining and conspiracy theories can’t bring down a good company.
Good job cashing in on the sensationalism, TC.
Sensationalism implies there’s nothing important behind the issue and homophobia (which it amounted to, accidentally or not) is worth getting riled up about. Granted it got overblown but that’s how the internet works. Lots of noice
Hodder’s explanation is cool adding to the information of what happened not just piling onto the emotional overdrive
I’m surprised it took outrage to get Amazon to shut down the thing. It’s a big pain in the ass for customers which can’t be a good thing. Last week a friend had a new release out and I couldn’t find it on Amazon.
Outrage aside, if people aren’t easily finding books they want there, they won’t be buying them there. That should be reason enough for any retailer to halt the system.
Oops.
noise, dammit. not noice
@Mary
You used some form of the word “ethical” several times. What does “ethical” mean to you? Who decides what is or is not ethical?
Ethical is that the cover shot for the second book is probably mercifully small.
Sounds like we’d better ban those evil Algorithms before they do some real harm.
Sometimes I wish TechCrunch actually read what they wrote from an audience perspective. It really does sound like she’s villifying the entire concept of “Algorithm”, which is pretty ridiculous.
I just think she’s calling for responsibility in their use. Shades of grey, right?
Vilifying algorithms by their nature (systemic, hard decisions, black and white) isn’t really productive. Imagine how radically easier it is to find books now that it was when you actually had to drive places and look on the shelves. That’s technology, and algorithms are inextricably linked.
Also, the discrimination made by an algorithm, whether it’s a reflection of its human parents or an accident, is more systematic than real human discrimination, and therefore more likely to be found – and when society cares, like it did with AmazonFail – they act. The action they took, in fact, would be wholly impossible without hundreds of algorithms working, fast, consistently, 24-7.
Algorithm – codified decision processes – are the core and currency of a technologically oriented business now. With a good enough instruction manual you can do anything. The value every individual or company creates is theirs. Mistakes happen.
Media people should interview “Ashlyn D” who sent the email in the first place. She or her boss have all the answers, assuming they’re not fired already.
“however that contradicts earlier email from them to authors stating that they were in the ‘adult’ category simply for including positive gay and lesbian themes in their works”
Can you show me where Amazon stated a book was considered adult because it contains positive gay/lesbian themes?
I see where they told the author his book was adult. And I see where his book is not adult, but rather includes gay themes, but I see nothing joining these two statements together into one like you’re stating here.
There is a huge difference.
So far, this amazonfail is a total non-story. Maybe that’s because I’m a developer and I know how easy it is to have something like this happen accidentally, and then you get the conspiracy theorists.
I mean, I apply Occam’s Razor here and it seems extremely unlikely that Amazon decided to ostracize a huge part of their customer base for no reason whatsoever.
I agree – this sounds like trying to turn molehill into a mountain. I don’t think Amazon is engaging in any political censorship – I think someone on the backend dev team goofed.
Really?
I did an experiment last week. I picked one of the m/m romance novels (Scarlet and the White Wolf by Kirby Crow) that had had its ranking removed and followed the ‘If you like this book you might like these’ links (links generated by Amazon, not by me). Every one of the related books also were ‘deranked’.
Then, I picked a m/f romance novel that still had its ranking (I thnk it was a Danielle Steele one), then looked at its ‘If you liked this…’ links. Funny, all of those books still had their ranking.
Small sample, sure. Still looked fishy to me, even to this straight woman.
Rankings for Kirby’s books seem to be back up now, at least.
:D
Great post. A difficult concept (for me anyway) very clearly explained. Thank you.
anyone who uses “napster” in an url needs to be flogged.
the reason why our culture sucks is b/c artist don’t get paid; they get ripped off; make something and try to sell it.
you sound bitter. learn to embrace the free market and web 2.0
sorry …test
Jesus H. Crist. Someone in France makes a mistake. Amazon cites a level 1 issue and pulls in folks from Easter Sunday festivities to fix it. And we get this tripe about fucking al´g&-rith-&m. Yes Amazon has a complex system. To suggest ethical (and/or moral) issues are at stake because of an inherent (prejudice) in the al´g&-rith-&m is just so stupid.
The problem is Amazon’s choice to classify all Gay positive literature as “adult” meaning: X-rated. This is offensive and is indeed a conscious choice made in the design of the system architecture – anyone who thinks it’s a mistake doesn’t understand how database driven design works. Whether it was conscious or unconscious the choice illustrates a bias — all coding choice have a Point of Viewe. I don’t see Mary saying that “algorithms are bad” as someone posted before. But rather that there is an ethical bias in choosing to classify all Gay literature as “adult” in nature. Doing so means automated processes can be applied that systematically filter and suppress results – in this case omitting results from sales ranking.
I don’t think that government should step and regulate Amazon — but I do think they should live by the rules applied to any other big company who wants to operate in a country that protects consumers (and employees) from descrimination. If Amazon wants to enjoy the revenues from selling gay positive literature – than it should make certain the fact of it is reflected in sales results. It’s just as if the NYTs best seller list was artificially skewed to omit African American writer’s work, which would be a descriminatory and biased (and I think illegal – and yes, un-ethical) form of supression.
Code isn’t neutral – it achieves the ends of it’s creators – and it has a POV.
PC leftist loonies …what wud u expect
I’m no lefty.
It’s a form of censorship. Even rightwing nut jobs would hate their sales results to be “misclassifed.”
Amazon is a private company. They have the right to censor what they want. If they want to ban all books by Republicans and left-handers, more power to them… but it will affect their image and sales… as will this.
Censorship necessarily involves force and the only entities with the legal use of force is government. It’s good to scrutinize companies… it’s part of why the free market is successful. But, don’t act like it carries the weight of Middle Age monarchs.
that is so one dimensional that it begs the question of how is it you have not a clue of the huge risk here should the gay community and the supporters of equality for all will do with this. brain dead.
Code has a point of view:
This is offensive and is indeed a conscious choice made in the design of the system architecture.
I am not sure whether it is a conscious decision or not. If Amazon is using concept clustering (ie, LSI – latent semantic indexing) which is more likely, then there is very little control over it. If one has to apply LSI to Amazon’s document space universe (text corpus), the terms :
- homosexual
- adult
- gay
will clustered into the same group, ie, they’re very close proximity to each other or to the center of the cluster. This process is not supervised, ie, the user or the designer of the system cannot control how clustering finds close proximity between items. This can be filtered via a rule after the retrieval stage (LSI or some other algorithms), but that might block other legitimate item recommendations where the concept “adult” is no way related to concept “homosexual”.
But this just proves the point. It was a human who made the choice, consciously or unconsciously, to place those terms in proximity and to weight those terms as something to filter out.
This glitch (which I have no doubt it was) simply revealed the unconscious biases of the programmers.
Wrong.
I see your point – if the classification algorithm was based on values entered by the programmer, then yes, it would be biased.
However, considering the size of Amazon’s catalog of products, this is more than likely done (as described above, if you read it) by an *unsupervised* algorithm, which means it gets the weightings from other data in the system, such as statistical info on it’s users.
Yes, wrong. It’s blindingly obvious why you’re wrong.
The whole reason you build an algorithm in the first place is that there’s more work to do than humans can possibly do. So you build a program, and give it the best logic you can.
You make it sound like “the algorithm flagged it because someone told the algorithm to flag it.” That’s not what an algorithm is. If a human could make all these choices, you wouldn’t need or have an algorithm!!
So get over it. Not everything an algorithm does represents a conscious choice by a human being.
Hm.
A. Somebody wrote that algorithm.
B. The algorithm was used to filter out “adult” books which might offend the larger public.
C. Gay themed books that contained the story of a stripper disappeared from the ranking (All I Could Bare).
D. Heterosexual themed books that contained the story of a stripper remained untouched (Candy Girl).
Blind as I am, I am still forced to conclude that either the algorithm contained an assumption which marked gay themed books as “adult” or the books themselves were tagged with a different descriptor than hetero themed books (which I highly doubt).
The entire point of the article is that algorithms reflect the unconscious biases of the human beings which create them.
Only someone who has an unreasonable belief in his own infallibility would fail to recognize that this is true.
Oh and additional exclamation points are not evidence of the truth of a statement.
Jack,
This is how it could become an “adult” theme.
Let us make the assumption that gay/lesbian people are more open-minded about sexuality than the average heterosexual person, This does not mean they go around sodomizing animals, but it takes certain convictions that sexuality is broadly defined (more than just liking the opposite sex) to identify yourself as gay/lesbian. Otherwise you would just deny it and dismiss it as conflicted feelings.
So with that assumption, is it a big leap to say that these people will, on average, purchase or at least view more sexually-related books and other products? I argue that they will because they are more open-minded about sexuality and will confront the subject more often than the average heterosexual person. Now some of this sexuality-related material could probably be classified as adult (this is without value judgment on my part…it is just a fact that it is classified as adult), and thus, if the algorithm takes into account adjacent associatons, then gay/lesbian could come to be associated with adult material. This outcome is correct in some cases and incorrect and others. The solution is refining the algorithm, not blindly pointing fingers.
It only made so much noise because of the gay fascists, who are looking for any excuse to attack somebody. This need for attention and fighting is getting very tiring. They just keep poisoning life for the rest us, who want to leave in peace.
Speaking as a gay fascist I would invite you to leave in peace whenever you want.
Greg, for someone who is very tired of gay fights, I suggest you leave too or get out of the closet!
Gay facist is an all time great oxymoron but bigots need to be outed and they need to be made aware that those of us who actually take equality seriously are going to get in their face and stay in their face and let others know who they are and where they work and who they hang with and where their kids are learning their hate. So where are you Grego cause you have a face i want to get into.
lol – yeah, a gay person can’t be a fascist… haha you’re funny…. and have quite the ego it seems.
Do you part son, go yell at all the people who don’t like you… i’m sure it’ll help
Oh I disagree. A gay person can be a fascist. For example Ernst Röhm, the original leader of the SA (Hitlers Brown Shirts) was undoubtedly gay.
Of course, he was closeted and it didn’t really work out that well for everybody, but he sure was a gay fascist.
People who use the term “gay fascist” need to be outed as bigots and idiots and, quite often, closeted self hating gays as well.
It’s unpleasant but in the long run it will cut down on gas chambers.
OK, Godwin’s Law, conversation over, I lose.
Thanks, Mary! Excellent points!
Facebooked and Tweeted it.
Great job – you looked at wikipedia and you now know what an ‘algorithm’ is!
This post is ridiculous, along with the outcry from the tech-ignorant masses of Twitter. As Falafulu Fisi said above, Amazon’s technology is going to be much more complex than making simple decisions based on book titles. Their algorithms are almost certainly based on data collected from user habits.
To put it even more simply – if a book is classified as adult, it’s more than likely because the average Amazon user thinks it is, and their systems picks up on their behaviour accordingly.
While I agree on the general point (algorithms can be biased), this does not apply at all to this specific case.
You assume that the misclassification of these products were a result of some automated process.. some algorithm that woke up last week and decided all gay themed books were “adult”. Do you have a source for this? It contradicts some actual journalism from the Seattle PI that says that the was misclassification was done manually: http://blog.sea...ives/166384.asp. This is consistent with a statement from Amazon PR saying it was a “ham-fisted cataloging error” (see http://www.lati...0,3536538.story).
As commenter anon said above: some dude in France screwed up and mistakenly labeled a bunch of books as adult products. There was no evil discriminatory algorithm at work here, just a silly data entry error.
Getting basic facts like this wrong a whole day after they are made public is pretty lame for a blog as prominent as TC.
“The bar for ethics in creating algorithms and classification systems should be very high.”
“I would suggest that the company, because of its position in the market and power over both authors and publishers, as well as users and the intellectual marketplace of idea, ought to be doing a complete and public review of their classification and algorithm assumptions. Publishers and authors should push for it, and so should users.”
In the same vein, I think that Google should immediately publish all the internal details of its PageRank algorithm, along with the associated data, for public review and content.
And Coke should immediately publish the formulae for its various products; who knows what might be hiding inside them?
And KFC must be forced to reveal the 11 herbs and spices that have made it such a success.
And the US Supreme Court should overturn the law of Gravity, since there was not proper review by Congress nor any signature by the President.
Excuse me whilst I go puke. This has to be one of the stupidest articles I have ever read.
Algorithms have no ethics; they are merely rules. The person who creates them (or the company that applies them) can do so ethically or unethically, but that is beside the point here. The fact is, if you don’t like the results, you’re going to complain, and you have every right to do so. You’re not even under any obligation to await a response, because its obvious that you already know everything there is to know (after all, you’ve looked up “algorithm” on Wikipedia).
Sad that TechCrunch puts up with tripe like this.
It must be nice to have Ron Paul and Ayn Rand to explain everything.
The amount of red herring being slung around here for feed at least of few nations of the world. This has nothing to do with any algorithms or other lame crap. This is not a private company either as some seem to contrive. Ask Coors beer if they want to comment on this if you want history on what bigoted statements and homophobia reap for any business that is caught with their head up their arse. This will cost billions and it could destroy this company in this time. To not understand that is ignorant, regardless of where you are on the topic.
If authors around the world pull their books from Amazon this week, what do you think Mr. Bezos will be doing this weekend? If accounts are closed, as i have closed mine, what do you think the response be? Will heads roll will the truth come out? If this is to go away, he better be working all night tonight trying to figure out how to get in front of this before he gets dragged behind.
So why would a book on “preventing homosexuality” NOT be categorized as “Adult” if this was an innocent algorithm glitch?
It seems to me positive images of gays was flagged while others were left untouched.
I wasn’t too bothered by this at first, having some background with data systems myself. But the more I see about it, the more disconcerting it becomes.
These books were “flagged” – it was not some magical innocent mistake. Items were marked in the Amazon database, and that is why the algorithm removed some titles and not others.
It’s not rocket science, and any attempt to make it sound innocent and “uncontrollable” are just ludicrous.
Because a book on preventing homosexuality doesn’t contain content that is graphically sexual. Many GLTB books do. That’s fine. That’s what people want. They want freedom to write what they wish. But you want a category? You want to write sexually explicit material? Then you will be put in a category defined as “adult” when someone gets around to categorizing in this fashion because they wish to cater to the apparently general public wish not to have explicit sexual material in their face all the time. That’s all. It’s a business decision.
Oh god. Here we go with the conspiracy theories again. So Mary can define algorithm. Big deal. Here theories based on her supposed knowledge of them are just more “let’s get my name in the news” crap of a sensationalistic journalist (but on TechCrunch that’s not a surprise).
And no, I’m not anti-gay for suggesting it’s not a conspiracy. I *AM* gay. I accept that whatever did happen at Amazon it was unintentional and that they are correcting it. They’ve always been very gay-friendly in the past, so it’s hard to believe they’d suddenly risk a PR nightmare like this and change course.
Let. It. Go. Already.
Your argument is long and detailed but keeps just repeating the same core claim: it can’t have been a glitch, because it’s an algorithm, and algorithms do what they’re designed to do.
As someone who has managed a large scale product catalog for a long time, I have some perspective on this. Algorithms are not pristine equations drafted into perfection, embodying the values of the parent company in that perfection. They’re spread out across many rules plugged into many systems that form an information chain. Yes, errors can occur. Sometimes an error that seems small in one part of the system can lead to major consequences down the line.
Add to this the inherent difficulty in defining what “adult” is, and I think you have all the makings of a glitch. I am sick and tired of people insisting that this was no glitch, and offering nothing more than their personal estimation that “it couldn’t have been.”
Consider what you’re implying. Amazon intended to screen out all lesbian and gay literature, because they think it’s dirty. Further, they thought they’d get away with it. BUT thanks to the superheroes who held them accountable with a Twitter shitstorm, they INSTANTLY backtracked, did a mea culpa, and fixed everything.
Your argument isn’t passing my smell test. And since you are accusing someone of persecuting gays and lesbians, I think the burden of proof is on you. You’ve proven nothing, except that you think of algorithms as magical and miraculous unicorns that can only reveal the hearts of their creators in perfect focus, and can never, ever, ever be off in any measure or just broken.
Have you ever built a search engine? Have you ever tried to organize a catalog of millions of things into a browseable structure? It’s not easy to get all the nuances right. Algorithms have points of view, etc etc, yes yes. But algorithms are also imperfect. I agree there is a difference between “sexual orientation” and “sexual chocolate,” but I defy you to build an algorithm that can differentiate it. Pornography is classically difficult to define, even moreso to a computer. So while apple-cheeked folks like you are SHOCKED that Amazon could ever let “sexual orientation” trigger an adult material trigger, I can easily see it happening. It’s not good. It’s fixable. But give them a break if they screw up their site for a day.
I’m frankly shocked by the naivety of your post, in which you claim that companies “with power in the marketplace” should permit outside review of their algorithms to keep them fair, and owe users transparency about how they all work. Heh. This shows in the first place that you think algorithms can incorporate mental constructs of yours like “fair,” and secondly that you believe some magical judgment panel could possibly agree on what “fair” is supposed to mean for a company like Amazon (which, while they’re waiting for the fairness committee to pass judgment on their algorithms, has, you know, jobs to create and profits to bring home).
If you want to read evil corporate conspiracy theories into everything, fine. I’m sick and tired of people doing that with everything under the sun all the time, as if mistakes never happen, and oppressive empires are waiting behind every bush to shovel us into the ovens at the drop of a hat.
Amazon’s adult filter has run afoul of gay/lesbian material in their catalog. Okay. They’ve been tuning it over a few months, it hasn’t been great, and they screwed it up big time last week. Some idiots in their customer service group sent out form letters that made it sound like Amazon considers gay literature “adult.”
I guess this company that sells tens of millions of products from thousands of merchants in hundreds of categories could possibly do a better job of it. I’m going to stop there, I wish more people could as well.
>Consider what you’re implying. Amazon intended to screen out all lesbian and gay literature, because they think it’s dirty. Further, they thought they’d get away with it.
No no no. She’s saying that the algorithm was constructed in such a way that it was able to erroneously flag a gay stripper memoir, but not a straight stripper memoir. It flagged gay children’s books. It flagged academic books about LGBT issues which only the truly depraved could consider erotically stimulating. It flagged books about sexuality and people with disabilities. And it didn’t flag books which think that gay people shouldn’t exist. The way the algorithm was constructed was biased towards judging books which assumed that gay people and other groups exist are more likely to contain sexual content–which is disturbing, but perhaps understandable, because that’s how society is biased.
>But algorithms are also imperfect. I agree there is a difference between “sexual orientation” and “sexual chocolate,” but I defy you to build an algorithm that can differentiate it.
Agreed. Nevertheless, hasn’t it been done before with Google SafeSearch, NetNanny, whatever? Which, granted, aren’t perfect (all of wikipedia being banned in the UK for a day was a major fail), but this doesn’t seem to have been a particular problem for them.
Once again, here’s the bitter truth you don’t want to hear: it’s not “society” or “amazon.com” that makes the GLBT category, and makes the GLBT category heavily sexualized. It’s GLBT lobbyists themselves that want a separate category, a “brand,” a separate section in bookstores, a recognized, separate identity (not mainstreamed) and it’s GLBT authors that create very sexualized works. It defies truth to say it is different.
So, live by the sword, die by the sword. It is where identity politics can lead you — to more discrimination, accidently or on purpose.
Here’s what’s all wrong with what you’re saying, Mary.
1. You are in denial that many GLBT books are very sexualized. You, and the authors of the books. That’s why in the New York Times, the college professor who wrote the book All I Could Bare: My Life in the Strip Clubs of Gay Washington, D.C could rage that his book disappeared and was flagged as “adult”. Er, why can’t a book like that be flagged as adults? All gay life isn’t about strip clubs, but *this book* is about gay strip clubs.
2. You are obdurate about understanding the public service motivation behind the decision to flag books in the first place (you appear ambivalent, or even to oppose the flagging of books). The point is that when a person types in “Washington, DC” into amazon.com “books,” they shouldn’t have to get a face full of “gay strip clubs in Washington, DC” — or any strip clubs for that matter — even if they are the most popular books — because that isn’t their general purpose when their type in a city name, I, like most users not in a niche would be looking for just general information about the tourist sites, history etc of this city.
I personally don’t feel any need whatsoever to have books “flagged adult”. If they show up in a search, so what?! But I can concede that amazon.com as an all-purpose book store being viewed by all ages and tastes would simply like to have that capacity to flag material as adult and not have it show up in general searches. Do you dispute that right?
3. No books were banned, and the fact that they were restored to the viewable list indicates that there was no discriminatory policy at work. There was merely a *screening out of sexualization* at work. You are bothered by that; you are also bothered by it when applied to GLBT. But if so much of GLBT content *is* sexualized, and it already is in a shelf of its own and category of its own in book-selling, *what* on earth do you expect, Mary? It will be flagged as adult. amazon.com said they were “ham-fisted” in how they did this — it was overkill. They are reversing it. But the original impulse — to screen out adult — and its adverse impact — on a body of literature that tends to the sexualized — is sound, and not discriminatory.
You seem to want to claim, by fishing for “Maurice” that the literature is not heavily sexualized. But there’s no denying it is.
4. We’d have to have a more scientific study of the 57,000 titles that were flagged. The glitch might have been that all gay literature of any kind, sexually explicit or not, got flagged as “adult”. That would be overkill. But given the high ratio of explicit sexual material in this category, *it would be reasonable, and not discriminatory*.
5. There’d be something to complain about if in fact E.M. Forster type of books that aren’t necessarily explicit in the way “Strip Clubs of DC” are explicit were now hidden from view (apparently amazon.com isn’t offering an adult filter to turn on and off?). But…they aren’t. In fact, now Strip Clubs is back, too. So it sounds as if in fact the flagging is a very liberal one.
Ultimately, unless you want us to believe that the making of any adult-flag algorithm at all is discriminatory inherently, or unless you want us to believe that gay literature isn’t heavily sexualized, you are wrong. And since many of us simply can’t believe either of these premises, your argument here isn’t convincing.
If you want to have a debate about how “no one should ever have an adult flag because it’s too hard to decide what’s sexualized, look at the huge range” — that’s fine. But that’s not what this is about.
Prokofy Neva, you miss a very important point here. I will grant (for the purpose of this argument) that many LGBT books are *sexualized* by their nature (and we can avoid the argument for the moment about how we define that) but so are a very, very large percentage of hetero-oriented books. One of the issues here is that comparatively equally sexualized books which were hetero- were *not* de-listed along with the LGBT books. In fact, books which were very hetero-sexualized were listed while LGBT which were not very sexualized were de-listed. Glitch or not, this is a serious bias and if it is the result of a conscious choice somewhere along the line needs to be addressed by Amazon, or they need to be prepared to face the consequences in PR and sales.
No. From everything we’ve seen about this, the issue is that the books *already flagged as gay* THEN got flagged as adult.
So the problem isn’t that hetereosexual sexualized books were skipped on purpose because the bias was against gays.
The problem is that the existing defined category of “gay” got flagged “adult”.
That check-off isn’t used on a lot of books. That’s why you could see so many hetero books that didn’t get the check-off — they weren’t seen as gay.
The marker “gay” when “adult” is layed over that then hits too many books that in fact aren’t so “adult” like “Maurice”.
My point, again, is this, in refutation of your claim of “bias”:
You can’t insist on a separate lifestyle/choice book category like “gay,” and then get pissed off it that then serves as the platform for making a mistake in a mass cataloguing effort It did; it was undone when the problem was seen. Trying to turn that into “discrimination” is silly, if you also want to retain the identity category.
Put another way: if there never was a marker or category called “gay” in the first place — which was a nod to a niche market — then there wouldn’t be a means to make a blanket error of this type. The “discrimination” comes in the original marking of the literature as “gay”.
One could argue, as it has been argued in other settings, that making special identity categories ultimately leads to discrimination. Here’s a textbook care of how that happened! If you don’t want collective punishment, don’t ask for a special identity.
If gay life were mainstreamed into all book categories as it may be someday and not a ghetto in the book store as a special identity, then this wouldn’t happen, eh?
What “PR consequences” is Amazon supposed to face for a) having marked all books about GLBT as “gay” in their system as a category when b) mistakes happen if adult is misapplied to this pre-existing category?
You want the label “gay” to be taken off so this will never happen again?
>What “PR consequences” is Amazon supposed to face for a) having marked all books about GLBT as “gay” in their system as a category when b) mistakes happen if adult is misapplied to this pre-existing category?
This is EXACTLY what Amazon said that they DIDN’T do. I quote from the pseudo-official statement: “It has been misreported that the issue was limited to Gay & Lesbian themed titles – in fact, it impacted 57,310 books in a number of broad categories such as Health, Mind & Body, Reproductive & Sexual Medicine, and Erotica.”
Perhaps you should better familiarize yourself with the facts of the issue before issuing these long rants.
Um, I know all that, and I’m as “familiarized” with the facts as you are.
You’re not getting the point.
The GLBT books are ALREADY categorized as GLBT in their system.
They flipped the switch to categorize books in a variety of categories as “adult”. One of those categories was GLBT. Anothe was “erotica”. Yet another was reproductive health or whatever. And so all those separate categories got another metalayer of data called “adult” that deprecated them from the search tag.
The problem is that not all GLBT is adult — but by blanketing that category with the “adult” tag on top of the GLBT tag they buried it. Same for erotica. But anti-gay books wouldn’t be hit because they didn’t fall into a category selected for adult to be added.
Thus, sure, it impacted 57,310 books in a number of categories, but because GLBT *as a categor of its own* was flagged, it was disproportionate.
That’s my understanding, and if I’m wrong, shoot me, but it sure sounds like that’s how it works from everything they have said on the record.
Of course, we are not peering right at the code and seeing whether it says “if GLBT AND sexual term in title MARK adult” or how it works.
It’s hilarious that you can both claim that in fact this mistake did NOT hit only gay books and also hit many others in a large grab-bag of 57,310 books, but still go on haranguing about “discrimination”.
You’ve also simply not understood what I wrote so I will say it again:
What “PR consequences” is Amazon supposed to face for a) having marked all books about GLBT as “gay” in their system as a category when b) mistakes happen if adult is misapplied to this pre-existing category?
You can’t dispute the books are marked as GLBT — they are! And they added adult to the entire batch.
It’s quite possible that not ALL GLBT books got flagged (surely there’d be more than 57,000 juts of them) but those that had “if GLBT AND erotica AND term in title” or something like that. We don’t have the algorithm to see how it works!
And again, my point is that the pre-existing identity category of GLBT is what set this up. If GLBT books weren’t in a ghetto, and were mainstreamed, it would be harder for so many GLBT books to get targeted in a mistake like this.
I’m sorry about the last comment I made. It was uncalled for. Here’s a more civilized response.
>It’s hilarious that you can both claim that in fact this mistake did NOT hit only gay books and also hit many others in a large grab-bag of 57,310 books, but still go on haranguing about “discrimination”.
Actually, I’m extremely doubtful about this 57,310 number quoted by Amazon. The way I see it, there’s a huge discrepancy between what was noticed by the #amazonfail crowd and what was claimed by Amazon. And I agree that it’s hilarious that Amazon could claim that there’s actually no discrepancy at all. What are these books that are categorized in Health and Mind/Body and Reproductive Technology that were deranked? Why didn’t we hear about it before? Why aren’t we hearing about it now? So ONE side of me says, the strong LGBT focus of the deranking observed over the weekend was in fact accurate, and Amazon’s claim is stretching the truth.
But the OTHER side of me says, let’s say Amazon was being honest, and there wasn’t any particularly strong LGBT focus, and that nobody noticed that all those other non-LGBT books were also deranked. In that case, we’re faced with something that you describe as follows:
>They flipped THE switch to categorize books in a VARIETY of categories as “adult”. One of those categories was GLBT. Anothe was “erotica”. Yet another was reproductive health or whatever.
Was it ONE switch which inappropriately flagged a VARIETY of categories of books as “adult”? This is what Amazon is claiming–not me, I repeat. Or were EACH of these categories (sexuality and disability, for example) individually and inappropriately flagged. If there is such a switch, can you tell me what it looks like?
And again, I apologize for my offensive tone. Debate is always healthy.
As I understand it, without being any more privy to their “algorithm” than anyone else, the “adult” flag went on top of other flags like “erotica” or “reproductive health”. And that means it turns up a variety of topics. And GLBT is “unfairly overrpreresented” because — gasp — a lot of GLBT literature is in fact explicitly sexual.
That’s something the crowd here doesn’t really want to look at or admit. For amazon, it’s ultimately not about “gay” but about “explicitly sexual”.
Why didn’t you hear about the other books? Because there is no networked, organized, angry lobby on Twitter instantly available for mass outrage in a category called “erotica books with nature photography that go over the edge into explicit”. But hey, call Flickr, maybe you can get a shitstorm going over there, too.
The least persuasive argument here is that “anti-gay books weren’t flagged”. But that’s because they aren’t sexually explicit. Let’s ask this question: does an anti-gay screed get a title GLBT in the first place? I doubt it, because what member of the GLBT constituency, going to that section of the bookstore to find their tribe and buy books, is going to want to have a nasty anti-gay book staring at them? The prudent bookstore would put that back in some other category like “sociology” or “health” or something.
It seems to me that those concerned about trying to screen out adult explicit content either consciously, with eyeballs, or automatically, with a set of tags like “GLBT” and “erotica” simply identified GLBT as a high-probability category for very explicit sexual material.
And in fact their action was legitimate because it *is* a category with a very high probability for sexually explicit material! There’s *nothing wrong* with that. It’s a free country. You can’t ban porn books anymore. You can’t ban even books with some graphic scenes in them. But if they re all in a group to in fact help customers find them better as a sales tactic and as a nod to identity, you can’t then later scream that they were “targeted” in a snafu like this.
If you are a business, you have the right to “front the merchandise”. Amazon.com is guilty only of “fronting the merchandise” away from sexually explicit material. They did this in a way that “ham-fistedly” hacked at swathes of material that wasn’t in fact sexually explicit or really deserving of the title “adult” but was only GLBT.
Again, the problem is a) a category called GLBT that in fact GLBT customers *want* b) a genre of literature that in fact contains a high degree of sexual material that in fact customers want, and that’s their right.
So to now bellow that amazon.com wishes, as part of an effort to push back explicit sexual content, to discriminate against gays, is to say that amazon deliberately decides to punish gays qua gays because they have literature that is too sexual. That’s false. The issue is that the literature isn’t gay, but sexual.
If you want to fight against amazon’s wish to front some merchandize and not others for the sake of what they believe to be the wishes of the general public, fight that fight, but you’ll find that amazon has likely done just that — catered to the wishes of the general public, as a business.
I completely agree. The mantra for Amazon is selection selection selection. More selection creates more choice and more customers and more sales and then they can buy more selection, and hey, that creates a virtuous cycle. But somewhere along the way, Amazon decided that they just didn’t want to be a business and sell stuff. They wanted to reduce selection so that they could let all those books sit around, raise costs and reduce sales. This whole thing makes so much sense. Silly Amazon, you forgot to be a business entity, you forgot that mantra you kept saying over and over and over for so many years, in a matter of weeks.
Apparently the Internet is full of idiots and TechCrunch thinks it’s OK to entertain one with a guest spot.
@Amazon : algorithms are gay.
So many people arguing about things they don’t know about (amazon’s codes)… yet so many people think they’re so right… ugh
Why is half the proof people use to accuse amazon of being biased based on some minimum wage customer service rep’s response?? You would think Bezos was the one who sent the reply…. but nope, it was just some chick in West Virgina or North Dakato and her email is being taken as God’s word on the subject and “the real truth” — that apparently they told the customer service rep and then went “DOH! Why would we tell her are secret plans?!”
Okay.. As an Ex-Amazonian.. I can assuredly say that this company is not biased in any way against sexual choice. This company is the most liberal company I have ever worked for. Probably 35% or more of this company lives alternative life styles…
As far as the code push bobo. Mistakes happen get over it. They do have an extensive check in system but that company doesn’t have a few lines of code – they have millions – Somebody probably did a change on the side to fix something and didn’t go through their change process system. I am sure they are either missing half of their but this morning or are packing boxes…
April 15, the worst day of the year in the US, and I haven’t even filed my taxes. Oh lawd! My way of procrastinating? Reading Mary Hodder’s article and the fascinating follow-up by techies, non-techies, homophobes, the “gay fascists” and people who have made visible the complexity of oppression, whether conscious or subconscious.
Technical ignoramus that I am, I never thought I would read anything on a site like TechCrunch, let alone get inspired by Mary’s article (sister, please distinguish between “its” and “it’s” in your otherwise enlightening article) and the fascinating correspondence which it spawned.
Are the head honchos at Amazon reading TechCrunch? If they are, will they recognize that they need to wake up and not demonize minorities by placing them in the dark basement of their ranking tower?
I will be watching another piece of the American culture war play out on the battle fields of algorithms. I hope that the Amazonians–those techie-bright and wealthy demi-gods at America’s largest online retailing company–will stop their peculiar policy of making invisible on their sales charts those authors and publishers who have the nerve to publish books in which gay men, lesbians, and transgendered people are actually shown as fellow human beings.
It seems is not ethical to criticize private individuals’ sexual orientation but it is ethical to criticize a private company’s choice of merchandising strategy. Such hypocrisy tends weakens the argument for gay rights. It makes some advocates of gay rights appear to be illogical and fascist.
It is completely ethical to question a company’s tactics and motives. It just so happens that they accusations being leveled at them in this case are fallacious.
Two ago, the Lambda Literary Foundation issued their Statement on the Amazon Controversy:
April 13, 2009–In response to the recent uproar over Amazon’s deranking of “adult” titles and its effect on LGBT books, Board President Christopher Rice has released this statement:
“Lambda Literary Foundation applauds the diligent work of writers, bloggers and activists in calling attention to this deeply distressing
turn of events. I have seen my first novel stripped of its sale ranking by this apparent computer glitch so I join other writers who are baffled to
the point of anger. I take great solace in the quick mobilization of our community in response to this apparent marginalization of LGBT books; the
grassroots power of the Internet has been placed on glorious display for all to see.
Over the next few days, we at Lambda Literary will be monitoring the situation very closely. Amazon is one of our nation’s largest general book retailers. In their commitment to creating and
sustaining technological advances in the publishing industry, they have laid claim to the future of book distribution. As such, they have a
pressing responsibility to create an unfettered exchange of stories and ideas.
If a quick and decisive response to this problem is not forthcoming within the next few days, we at Lambda Literary look forward to leading a
sustained and impassioned dialogue on this issue, which will seek to harness the energies that have been released by our community’s admirable response.”
No “assumption” is necessary in the algorithm. Do you think Amazon goes through each book tagging it manually?
If they do it automatically, then all that’s needed is simply statistical correlation. I.e. IF the people who read homosexual books tend also to read adult books to a higher extent than the people who read Winnie the Pooh, then an automatic classification system will group adult and homosexual categories together.
Great work and great article. I used to be a programmer and it was useful to have an article written by someone who understands how IT works. Again, great work. I’ll never use amazon again.
Hodder’s explanation = FAIL. First, as others have said, you can’t take a low-level CSR’s response to a customer as an informed and general statement of management policy. (And Hodden misrepresents what the CSR said to boot, which I find astonishing, since she quotes the whole thing, and sort of embarrassing too.)
Second, to suggest that there are only two ways to categorize books about homosexuality (as negative or positive) and that therefore Amazon must have already set out to distinguish between them ignores that there are obviously dozens of significant keywords in play. When you use the search terms “gay spirtuality” for instance, anti-gay religious books do not pop up, although some of those books are highly ranked. Does that make sense with the kind of simplistic duality Hodder infers? No.
Amazon acknowledges that books can be tagged as adult. I don’t see what is implausible about the French guy explanation. Looks like hysteria to me at this point.
Looks to me like the customer service rep gave exactly the correct answer. The books were flagged as adult, and adult books are excluded. That’s the rule. How would she know if a book were incorrectly flagged as adult?
As a result of the conspiracy-theory nonsense, I will now buy as much as possible from Amazon.
This is laughable. “morale responsibility”? I can’t even believe that someone would publish this drivel.
Amazon is a company that delivers on what is wanted and needed. They do it better than anyone else. That is why they are successful. They are not who they are because they are fair or not fair.
Amazon does not have a policy of attacking anyone. Any organization that has 1000’s of employees, each with their own beliefs and morality, will occasionally have a problem with one or some of them acting in their own self interest. When enough people complained, Amazon looked into it and corrected the problem. A perfectly fair and reasonable response.
Why don’t you take your “morale responsibility” and protest the Muslims that stone their wives and daughters, the Iranians that behead gays for being gay, or the Chinese for their civil rights violations.
Please climb down from your ivory tower, pull up your big girl panties, and worry about something that IS actually a problem.
I have always thought of websites as a ’store’. When you own your own store, you decide what isles to put the milk in, and the eggs.
It is usually done with the intent of selling something better.
Algorithms are man-made and a representation of the personality of the creator(s). I doubt, seriously, that this is done with personal intent, but more so as a business decision to see what can sell more.
The current state of the web-customer is that they view online activity as potentially evil, simply because there is no person reassuring them that all is ok, in person.
Sometimes fear, the way one is treated, his/her insecurities, or ability to reason emotionally & logically…makes people irrational, and causes them to believe that technology is aimed at harming them. Yet….If I cut off a frogs leg, then told it to jump, and it did not move, could I not deduce from this that maybe the frog was deaf, or was ignoring me deliberately? Get real.
If you dont like a store’s policy, maybe it is best to shop elsewhere. It does not make it a personal attack if McDonald’s filters out selling onion rings, metaphorically speaking.
Steven
Voice123.com
The issue with #AmazonFail isn’t that a French Employee pressed the wrong button or could affect the system by changing “false” to “true” in filtering certain “adult” classified items, it’s that Amazon’s system has assumptions such as: sexual orientation is part of “adult”. And “gay” is part of “adult.” In other words, #AmazonFail is about the subconscious assumptions of people built into algorithms and classification that contain discriminatory ideas.
This seems a little unfair to me. Without actually looking at Amazon’s source code, it’s hard to claim that there were any offensive “subconscious assumptions” built into it.
I propose this question for your consideration: Could a similar glitch have branded other categories of media as “adult?” For example, could this have happened with children’s books rather than LGBT stuff?
Even if Amazon did have an “adult” flag on some materials and that is what caused an inadvertent change in classification, the false assumption is being made that their internal “adult” flag has a one-to-one correspondence with what responsible, impartial citizens would view as “adult” material. I work on a system that has a “secure customer” flag and, surprise, some people with the flag are neither secure nor customers.
This is just another example of why Amazon has become a hated company by many authors. The last big thing was removing the buy buttons of authors who’d published with various small presses. Authors and publishers just don’t trust Amazon’s intentions.
John, I agree with you that there are more pressing moral issues like NOT supporting gay men who get executed in Iran, or NOT supporting women in Pakistan and Afghanistan who get raped and stoned (although AMNESTY INTERNATIONAL is doing its best to help), than getting upset about Amazon’s policy of apparently discriminating against books with lesbian, gay, bisexual and/or transgendered subjects.
On the other hand, chances are slim that we can influence radical mullahs in Iran or Taliban and other radical Islamicists in Afghanistan and Pakistan, but we have a chance to send a wake-up call to the world’s largest book distribution company to investigate claims of discrimination and, if found to be true, to rectify the situation.
A note of caution: Members of a majority–say white or heterosexual or middle class or Christian–rarely seem to see, let alone experience, the pain that those can experience who are not members of those privileged groups.
Reading some of the comments on this string I get the impression that myopia has created some strange heterosexist views, hidden behind a “free market” ideology, a modern American equivalent to the classical “why don’t they eat cake?” ignorance.
One does not have to be gay, lesbian, bisexual, or transgendered (or “questioning,” for crying out loud) to see that the time has come to recognize some new realities and, as a result, perhaps change the algorithms of the mind, not only those at the Amazon headquarters, but those deeply seated algorithms in our own heads.
What do you think?
“pain” — and that would be…having lost search rankings for a couple of days?
I think you *can* care about other causes that are really about a lot more pain, like the violent persecution of gays in some countries, and even work toward legalization of gay marriage in the U.S.
This was a self-indulgent and hysterical campaign that spread as fast and far as it did precisely because it was limousine liberalism on the Internet, click, RT, click, express outrage, done.
I’m not going to lobby for any new “algorithms of the mind” when you simply haven’t persuaded me that there was any discriminatory algorithm in the first place. You haven’t responded to any of my arguments.
I doubt that the pain Socrates Socratis was about the page rankings being removed, but instead was about the criticism of the reaction by those who protested. Or, wait. I’m not really wording it correctly . . .
1.) Someone who is beaten time and time again might start flinching when someone walks in the room. There have been a few GLBTQ victories lately, but they feel somewhat precarious, even as groups opposing gay marriage vow to file referendums and run commercials about impending storms. Knee-jerk? Probably, in hindsight. But underneath that reaction is the realization that there is a problem with assumptions or attitudes that appeared to be playing out here (different rules for gay and het stuff). No matter how liberal Amazon’s ethos seems to be, the truth is that Amazon is a behemoth corporation – and people are a little raw about anything that looks like “the majority” (ever-shrinking as it may appear).
2.) Stripping away *everything* around this, say you are a kid who feels like no one understands what he’s going through, goes to Amazon, searches for a book (what kid goes to a library website anymore? AMZN *is* kinda the default). Only results that come up talk about overcoming or preventing homosexuality. There IS resulting pain from that, and it’s not about the page-ranks.
3.) Those belittling that reaction and saying “it’s no big deal” legitimize, to a certain extent, this very real fear — a group of people who sometimes shout very loudly as a result of feeling unheard and barely represented, being told they overreacted. What would have happened if, for some bizarre reason, an algorithm worked to have every book on a search for “Religion” end up being about debunking of same? How many heads would have rolled?
4.) Attitudes on the whole are changing and the whole country, no matter what side, is pretty raw about it. So while people may have overreacted, and I certainly think that the marriage of algorithms and assumptions needs a little work, and I also think that outside review of these algorithms IS heavy-handed, I’d say cut people some slack.
Oh, stop whining.
If you want to work on gay marriage rights, work on that directly, don’t have an infantile placebo spasm banging on a corporation to get attention.
Work on persuading your fellow Americans who need persuading. amazon.com is not the problem.
You keep repeating over and over that there *was* discrimination deliberately and therefore you are “entitled” to some kind of hurt, and some kind of compensation.
But there wasn’t any deliberate discrimination. So you are not entitled to anything.
There is no evidence that there were any deliberate algorithms to deliberately flag gays as adult inappropriately. So why are you continuing to bang on this? Why is anybody?
Prokofy, not sure why it’s not letting me reply directly to you.
Trust me, many of those (including me) that were worried by what happened DO work directly on gay marriage rights.
I am not repeating that there was deliberate discrimination. I’m saying that though it now appears there wasn’t, there THEN appeared there was. To say that people should just get over it once again minimizes the effect of these attitudes.
I’m not sure where I said anyone was entitled to compensation. In fact, I’m sure I didn’t. I think I maybe suggested cutting people a little slack, taking a minute to realize the reason (not excuse, but reason) for the freak-out. Then I gave reasons for that suggestion, not to suggest that anyone was entitled to pain, but that it might have been an unfortunate side-effect of a mistake that, despite being unintentional, still had consequences — just like mistakes (mostly unintentional, really) tend to do.
In fact, my point really was that yes, there are fellow Americans that need persuading. I suppose since I’m talking about it here instead of elsewhere it might seem like I’m banging on Amazon or specially on the issue with Amazon, but my point was more that this might be an opportune time to analyze assumptions both pre- and post-deluge.
Geez. Who’s spasming?
(And at any rate, in looking at your blog post on the subject I see a better encapsulation of what I’m trying to say, about the free-floating rage, and you’re right, it does need to be channeled. And I hope this episode will make others realize the same thing.)
Kristin,
The cause is never served by falsehoods. Indeed, the argumentation that the religious right often makes in the YouTubes and the blogs is that if you grant gay rights, you will lose your own civil rights and freedom of choice. So they shouldn’t be given quarter there, and it should never seem that fighting for gay rights involves hysterical persecution of people perceived to somehow slight gays — when in fact they did not, and it is false to charge them in that manner.
You keep saying this: “there THEN appeared there was.” But, there wasn’t.
As much as I support gay rights, I refuse to turn #amazonfail, even in the “morning after the morning after quarterbacking” into a gay right issue. It isn’t one. In fact, the infantile and whiney and entitlement-happy behaviour of so many in the Twitter mob over this, screeching and demanding amazon’s head, was just wrong, wrong, wrong, and only feeds into the stereotypes about what gay rights are. They should always be about equality and mainstreaming, and not about special privileges and suppressing others’ rights.
As much as he is wrong about many other things, Clay Shirky impressed me by being able to accurately analyze and critique is very own mob behaviour on this in giving in to the seduction of mass self-righteousness. He is spot on:
http://www.shir...-of-amazonfail/
Sorry, but I don’t understand why this is being discussed as anything to do with the search algorithm; surely it’s about how Amazon subdivide their catalogue?
It seems to me that what’s happened is this. A decision was made to strip all books in the erotica subcategories of sales ranks. The process by which this was done was also applied to the LBGT subcategories. The relevant question therefore is how did that second event come about and was it the result of a conscious deliberate management decision? Personally I doubt Amazon would deliberately do something that stupid, so I’d guess human error in applying policy sent down from above.
The key claim “…underlying assumptions that drove their classification and algorithm system to be built to filter “gay” into “adult”” just doesn’t stand up.
I still don’t understand why Amazon needs an “adult” tag in the first place. You have to be an adult to have a credit card, ergo you have to be an adult to use Amazon. There is no reason for any censorship. Granted, minors can browse Amazon, but I don’t see how seeing the titles (& sometimes covers) is going to be a problem. In fact, I’m not a big fan of censorship even for minors, but that’s a separate argument.
Certainly a lot of back and forth about coding and algorithms on this post, but I would like to suggest that Amazon’s big mistake in this issue is their lack of PR savvy. The responsible thing would have been to say first, “We have heard complaints and we’re investigating the issue.”
Then they could say, “We see that there are some challenges in how our algorithms are categorizing books. For now, we will restore previous settings and develop a plan to address this.”
Finally, the could say, “We have investigated the matter and have found out the following and here’s what we’re doing about it.”
The lack of clear communication on this issue by Amazon is compounding the problem. Regrettable that they don’t appear to have (mini) crisis management in place, or at least a proactive PR department. Their customers and their authors deserve better.