Amazon Freaks Out About Sock Puppet Reviews And Deletes A Bunch Of Real Reviews

from the collateral-damage dept

For a while now, there has been a bit of a kerfuffle at Amazon over so called "sock puppet reviews" or reviews purchased by an author to help pad their books' rankings. We hadn't been covering any of it because, frankly, it was a non-story. There never was a threat to the publishing industry and it was always questionable how widespread the problem really was. Additionally, the idea that a writer would have to pay to get reviews was just a sign that those writers held no real confidence in their work.

Unfortunately, Amazon took these complaints a little too seriously. It would seem that those complaining were loud enough that Amazon heard them and did a couple of things to tackle the non-issue. First it revised its rules for review writing. to make such purchased reviews against the rules. Then it removed a bunch of reviews seemingly at random. Joe Konrath shares his experience upon reading about this:

I've been buried in a book deadline for all of October, and haven't been paying much attention to anything else. When I finally took some time to catch up reading email, I noticed I had many authors (more than twenty) contacting me because their Amazon reviews were disappearing. Some were the ones they wrote. Some were for their books. One author told me that reviews her fans had written--fans that were completely unknown to her--had been deleted.

I took a look at the reviews I'd written, and saw more than fifty of them had been removed, namely reviews I did of my peers. I don't read reviews people give me, but I do keep track of numbers and averages, and I've also lost a fair amount of reviews.

Why did Amazon go nuts deleting reviews? Well, Konrath assumes, based on his responses from Amazon, that this was the result of a new automated sock puppet detection program. Apparently, it works in much the same way as Google's ContentID: flag anything and everything and see what sticks. Actually, no. This is way worse than ContentID. At least ContentID has some kind of -- admittedly weak -- notification, human review and appeals process. That is entirely absent from Amazon's deletion program, as Konrath explains in his letter to Amazon.

My reviews followed all of Amazon's guidelines, and had received hundreds of helpful votes. They informed customers, and they helped sell books. They represented a significant time investment on my part, and they were honest and accurate and fully disclosed my relationships with the author I reviewed if I happened to know them. And these reviews were deleted without warning or explanation.

Next, in his letter, he explains just why Amazon's actions were the wrong thing to do. Primarily because this action harmed more authors than sock puppet reviews ever did.

Obviously Amazon can do whatever it wants to on its site. It isn't up to me to dictate policy. It's your company, your rules, and I fully respect that. But I believe Jeff Bezos is very much about treating customers fairly, and I've heard it said many times that Amazon considers its authors to be valuable customers. So you should know that I'm just one of dozens of authors who are saddened by this, and those are just the ones who have emailed me.

The community you're trying hard to nurture is upset by your actions. They feel those actions are unwarranted and harmful.

Please express our disappointment in Amazon to anyone who needs hear it, and let them know I'll be blogging about it. People are seriously disappointed in how Amazon handled this. It was a knee-jerk, inappropriate reaction to a ridiculous case of unjustified moral panic, and a Big Fail.

Admittedly, this act by Amazon was in response to a number of authors who complained about the problem. However, as I wrote above, it was a problem of egos, not actual harm to any specific authors or group of authors -- or as Konrath put it, an unjustified moral panic. Authors freaked out over news stories of people being paid to write reviews and it ballooned from there. And just like every other moral panic before it, this one did tons of unnecessary collateral damage.

So not only do a bunch of legitimate reviews just up and disappear, there is also further damage to Amazon and the authors it works with. Readers will be less likely to write thoughtful and meaningful reviews in the future. If your review that you spent an hour writing could just up and disappear, why bother? Is this really what Amazon and these authors want -- people less willing to review books they read? That would seem to be a far worse situation than an unconfirmed number of sock puppet reviews.

Filed Under: books, reviews, sock puppets
Companies: amazon

Manufacturer Of Buggy 'Kobo Touch' E-Reader Manages Customer Complaints By... Hiding All Online Reviews

from the there's-no-hole-on-earth-big-enough-to-bury-The-Internet dept

Quick quiz on crisis management, internet-style:

Exercise 1: The product you've just introduced is a buggy mess, short on content and backed with terrible customer service. What do you do?

a.) Bite the bullet and start handing out refunds
b.) Start patching like hell and fire your current Customer Service team
c.) Drain all bank accounts and reorganize under the name Net Sortie, LLC.
d.) Whistle nonchalantly while sweeping bad reviews under the rug

If you answered "d," then congratulations! You've lost the battle and the war!

No matter how many companies line up to play the "I'd Like to Lose at the Internet" game and walk away empty shells of their former selves, there's always another player ready to step up and take a swing at wishing its problems into the Google Cache cornfield.

This week's contender is Rakuten and its Kobo Touch Reader. Billed as sort of a preemptive strike against the expected arrival of Amazon's Kindle in Japan, the Kobo began shipping last week. That's the last of the good news.

Rakuten launched the Kobo Touch in Japan with the expectation that they would dominate their home market. They are native to the country, and Rakuten does have a sizable retail presence there. Given their technical and CS resources, you’d think they would have been able to pull this off.

Unfortunately, it now looks like Rakuten has paved the way for Amazon to dominate yet another ebook market.This launch is rapidly turning into a debacle and it’s going to damage Rakuten’s reputation. And according to some of the tweets I’ve seen (in Japanese) it already is.

So, what went wrong? Well, many, many things. First off, while the firmware was solid, the desktop software was a disaster. If installation failed on the PC, a rather common situation according to the reviews, it pretty much made the Kobo Reader useless. Secondly, Rakuten's promotional work pointed towards 30,000 titles being available at launch. Instead, there were 18,894 titles and, as is pointed out in the comments, many of those were public domain. Last, but not least, purchasers now holding a shiny brick were treated badly by Rakuten's customer service.

Rakuten was understandably perturbed by this failed launch and decided the best course of action would be to pretend it just wasn't happening.

It’s been just under a week since the Kobo Touch started shipping in Japan, and things are going so well that Rakuten has removed from their website all the reviews of the Kobo Touch.

No, seriously, all of the reviews are down – both good and bad. I suppose there were too many people writing things like “I’m going to buy a Kindle” and that upset someone at Rakuten.

For a tech company, you'd think Rakuten would be a bit more familiar with how this "The Internet" works. You can't just pull the electronic wool over everyone's eyes and hope to sneak away undetected. The Internet never forgets. And even if it could, there's always a helpful person or two willing to remind it where all that stuff is stashed.

Luckily one of Rakuten’s potential customers tipped me to the story, including giving me a link to a blog that had collected responses and a screenshot of the review page before Raskuten removed it. That’s why I can show you things like this:

If you can’t see the image above, it says that the Kobo Touch has a 3 star rating largely due to the vast number of 1 star reviews.

And more help arrives:

Update: My source found a Google Cache page of reviews. Thanks, Bibo!

This response is so wrong and yet so common. Donna Barstow, law schools, the freakin' medical community. Pretty much anyone who's ever heard the term Streisand Effect whispered in their general direction has attempted to delete damning content, either of their own or created by others, only to find it resurrected in Google's Cache or the Internet Archive's Wayback Machine. It's definitely a knee-jerk response, but it seems to have a steep learning curve attached to it. Kobo screwed up and then doubled down by hiding the reviews. What does that say about the company and its future relationship with its customers?

Folks, they took the reviews down from the website so new customers wouldn’t be warned about the many problems. I want you to look past the fact they did it and think about how customers will feel once they discover the deception. That is what will make this a major debacle and not merely an embarrassment for Rakuten.

A bad launch could be recovered from. This is closer to being a systematic effort to lie to their customers. Okay, eventually people will forgive Rakuten, but in the short run this debacle could drive readers to Amazon.

Rakuten, by botching its launch, hurt itself a little. By covering it up, it did a ton of self-inflicted damage. You can't just flip "trust" on and off like a light switch. It's earned. And if it wanted to take on Amazon, it couldn't afford a mistake of this magnitude.

Filed Under: competition, japan, kindle, kobo, kobo touch, reviews
Companies: rakuten

Gym Allegedly Threatens To Call Police On Blogger For Blogging That Its Prices Were A Bit High

from the good-luck-there dept

Every time you think that there have been enough examples of the Streisand Effect that it should stop people from trying to censor perfectly legitimate content, you're just setting yourself up to be surprised by the next such attempt -- often more bizarre than the previous one. The latest one comes to us via the Consumerist, and it involves the story of a new gym in Cleveland, called BarreCleveland, that has apparently threatened to call the police on a blogger because she though the price was a bit steep. But the really odd thing? The original post was generally positive. We're used to seeing people flip out about people saying negative things about them, but a mostly positive review? Now that's something special.

Alana Munro's original blog post did discuss the fact that she thought the price was too high, and that she normally likes a workout that generates more of a sweat, but on the whole it doesn't seem that negative at all. She noted that if it was a little closer to where she lived (and the prices were lower) she'd likely go back. However, somewhere along the way that got interpreted as being a dig by someone who ran the place, and they first got into a bit of a Facebook and Twitter argument with Alana and some others.

And then.... someone with the same last name as the owner posted a comment on the original blog post, accusing Alana of "stealing" the class she attended:

Just stop the posting about Barre Cleveland and take down all the existing posts. We know that you stole the class and we can pursue legal action against you for that and that is why it is ridiculous that you complain about a price when you never paid for the class. You were never given a discount code by Barre Cleveland and somehow you used that to enter the studio. I am sending you this message to politely ask that you remove all the content about Barre Cleveland from your blog and twitter and we will not get the Beachwood Police involved on this theft of services.

Alana says she did use a discount code -- it was one that was being passed around widely to help promote the gym in the first place. Now the gym may argue that it only meant the code to be used by people it gave it too, but then it should have implemented one-time codes. Instead, the code worked, and Alana has a nice receipt posted to her blog "thanking" her for signing up. Her complaint wasn't that she paid too much for that one class, but that the general price might be too much for her to keep going back.

Either way, the whole thing has blown up and received lots of attention. BarreCleveland has scrubbed the back-and-forth tweets from its Twitter feed. But it certainly looks like it already went way too far. Engaging with people who criticize you online is one thing -- and can be quite useful. Threatening to call the police on them and demanding they erase what they've written about you is going way too far, and the internet (hello Streisand Effect) doesn't take kindly to such things.

Filed Under: alana munro, censorship, cleveland, reviews, streisand effect
Companies: barrecleveland

Real Estate Firms Slipping 'Non-Disparagement' Clauses Into Rental Agreements To Stifle Online Criticism

from the makes-you-wonder-what-they're-afraid-of dept

Rob Hyndman points us to travel guide guru Arthur Frommer speaking out about a ridiculous trend from some real estate firms that do vacation rentals: slipping a non-disparagement clause into contracts that effectively forbid you from writing a bad review about them online. Frommer's discussion is built off of an article Christopher Elliott wrote for the Washington Post, highlighting the trend:

But don’t go looking for the Dorows’ opinion on the Web. Within a few days of posting it, they received a letter from their vacation rental agency.

“It has come to our attention that you have written an unauthorized review regarding your stay at a home managed by Progressive Management Concepts,” it said. “If this review is published by VRBO.com, you will be in violation of the confidentiality clause of the rental contract you agreed to when you made your reservation.”

When the Dorows refused to remove the review from VRBO.com, the site through which they’d found the rental, Progressive Management promptly charged $500 to their credit card.

If this kind of thing sounds familiar, you may be remembering how some doctors and dentists have done similar things (though using an even more nefarious method of demanding assignment of the copyright on any such review).

Frommer, reasonably, finds this behavior deplorable, and notes that this is only likely to expand to lots of other things that get reviewed online, such as hotel stays. As he notes, since properties like this already have your credit card on file, it's a real risk that they can then just try to "charge" you for writing a bad review. Frommer also disputes the arguments of those in favor of such clauses by noting their argument is basically that they don't like free speech:

Now the various people who defend these clauses, base their arguments on all sorts of horrid, potential and hypothetical threats. They claim there are a lot of people who, upon checking out, threaten the rental property with a negative review unless they are given a retroactive discount on the rental. They claim, in effect, that vacation renters are blackmailing them.

In my view, simply to state that argument is to refute it. You could justify a great many denials of our First Amendment rights of free speech with scary hypotheticals like that.

Of course, the best response to something like this is to recognize that if you put such a clause into a contract, it means that you're hiding something, and aren't confident enough in the quality of your property that you can handle criticism. It should be an automatic disqualifier for a renter. Unfortunately, though, as the article notes, sometimes you don't even get to see the terms until after you agree. While I'd argue that makes such terms unenforceable, it does make things more complicated. At the very least, though, it should allow people to post a review that merely notes the fact that such a clause is in the contract, and how that suggests no confidence in the quality of the property...

Filed Under: disparagement, real estate, reviews, vacation rental

Dentist Who 'Invoiced' Patient For Negative Reviews, Getting Slammed On Yelp

from the these-things-have-a-way-of-coming-back dept

You may recall that, yesterday, we wrote about the class action lawsuit filed against dentist Stacy Makhnevich. Makhnevich used ethically and legally dubious forms from the organization Medical Justice, to demand the future copyrights on any reviews a patient might write about her. Then, she used the DMCA process to try to take down negative reviews on Yelp and DoctorBase. When that didn't work, she threatened the patient, Robert Lee, with a lawsuit, and started sending him invoices for infringement, at $100/day. None of this addressed Lee's original complaint -- that Makhnevich failed to submit the documents he needed to get reimbursed from his insurance company for an expensive procedure.

Of course, as with any typical Streisand Effect situation, all this ended up doing is leading to a hell of a lot more attention to the situation and the negative comments. But, these days, things can go even further than just driving more attention to content someone wanted disappeared. It can lead to even further backlash -- especially on sites involving reviews -- as we've seen with authors who get dinged for questionable actions. If you go take a look at Yelp's page for Stacy Makhnevich the one star reviews are flowing in... many of them calling her out for what she did. Oh, and Robert Lee's review, which kicked this whole mess off... is Yelp's "featured" review at the top of the page. Her current total review rating is at a star and a half. It used to be much, much better. Some of the reviews are entertaining. I liked this one, which notes "I heard you have to bring your lawyer with you to the dentist's office." For future reference, if you're that concerned with your online reviews, perhaps just do the best you can and respond to customer complaints promptly. Trying to whitewash complaints seems likely to backfire in big, bad way.

Filed Under: class action, copyright, dentists, dmca, doctors, reviews, stacy makhnevich
Companies: medical justice

Dentist Forced Patient To Sign Away Future Copyright On Any Online Review; Then Billed Him $100/Day For Negative Reviews

from the copyright-abuse dept

We've talked about Medical Justice a few times. This is the highly questionable outfit that gives doctors and dentists forms that these medical professionals then require patients to sign in order to get treated. The forms require the patients to pre-emptively hand over the copyright on any future reviews they might write about that medical professional. The idea is that if the doctor or dentist doesn't like the review, they can then just use the DMCA to take it down, claiming that the review infringes on their copyright. Yes, this is incredibly sleazy. This clearly has nothing to do with copyright, but rather it's a use of copyright law to try to censor criticism. There are both ethical and legal problems with this... and the legal problems are about to be discussed in court.

Public Citizen has filed a class action lawsuit against a New York dentist, Stacy Makhnevich, who not only used the Medical Justice forms, but then sent one of her patients invoices, supposedly billing him $100 per day for having posted comments about her online. As Paul Alan Levy explains:

Our individual client, Robert Lee, had a bad experience, not with Makhnevich’s dental work, but with her billing and her failure to submit the documents he needed to get reimbursed by insurance.  After his repeated efforts to get her office to do what they were supposed to do, he posted complaints on Yelp and on DoctorBase.  Makhnevich threatened to sue him over the posts, and sent DMCA takedowns, but no doubt to her surprise, not only did the patient not remove his comments, but both Yelp and DoctorBase defied the threat of infringement liability, telling Makhnevich that they regarded her agreement with the patient as illegal.  Undeterred, Makhnevich sent Lee invoices purporting to bill him $100 per day for the continued copyright infringement.  Makhnevich also hired a lawyer who sent additional threats of litigation, but rather than continue to wait to be sued, Lee has now filed suit for a judgment declaring the agreement void, an injunction preventing Makhnevich from imposing the agreement on other patients, and a notice to all Makhnevich patients informing them that they are no longer restrained by the agreement.

A few interesting things come out in the lawsuit. First, the fact that both Yelp and DoctorBase defied the DMCA takedowns. Both companies deserve kudos for that. Standing up to bogus DMCA takedowns is pretty rare these days, because the risk of getting roped into a costly lawsuit is just too high. In this case, the fact that there had been so much news about Medical Justice and it's questionable concept, and both Yelp and DoctorBase were aware of this earlier, certainly helped. Still, standing up to such threats deserves praise.

Second, the lawsuit notes that not only is this copyright abuse, but the DMCA takedown notices appear to violate HIPAA -- the federal Health Insurance Portability and Accountability Act -- which is supposed to guarantee privacy for patient info.

In September, 2011, on the letterhead of Aster Dental, a member of Dr. Makhnevich’s staff sent takedown letters under the Digital Millennium Copyright Act (“DMCA”), 17 U.S.C. § 512(c)(3), to Yelp and to DoctorBase, asserting that Dr. Makhnevich owned the copyright in Lee’s Commentary pursuant to the Agreement. Defendants warned Yelp and DoctorBase that, if they did not remove the posts immediately, they would lose the immunity that the DMCA otherwise provides Internet Service Providers against monetary liability for copyright infringement. In violation of HIPAA, these takedown notices disclosed plaintiff’s height, weight and birth date, as well as his picture and his home address

The lawsuit also makes the claim that these comments, even if the copyright on them has been assigned, would still be protected as fair use. But, more importantly, it argues that these agreements in the first place constitute "copyright misuse" noting that the agreements:

constitute unclean hands and, with respect to such purported acquisition and assertion, constitute copyright misuse in light of the means by which defendants purportedly acquire the copyrights and because the purpose of such acquisition and assertion of copyright is to suppress truthful commentary concerning defendants and matters of public concern.

No matter what, this should be an interesting lawsuit to follow. As Levy explains, beyond the legal arguments, there is an important public policy aspect to this lawsuit:

The purpose of copyright law is to encourage creative expression by providing a temporary monopoly (sadly, less and less temporary) that enables those whose expression is marketable to reap financial rewards for their work. At the same time, copyright law avoids giving any monopoly on facts or ideas. Agreements like the one at stake in the Makhnevich case turn copyright law on its head by taking advantage of the fact that, as a practical matter, ideas and facts are articulated through copyrightable expression; hence anything that a patient writes about a doctor or dentist is likely to have sufficient originality to be copyrighted. The Medical Injustice agreements allow professionals who use them to suppress the underlying opinions and facts, not to reap financial rewards from the expression and not to encourage further creativity. This is a misuse of copyright law and in our view it needs to be stopped.

However, from a future policy perspective, it appears that this lawsuit has already been a win. Within a day of the lawsuit being filed, Medical Justice has announced that it's retiring the agreement... and that it probably should have earlier. It also claims that it's telling doctors to stop using them. We'll see if that actually happens... and if anyone else jumps into the fray instead.

Filed Under: class action, copyright, dentists, dmca, doctors, reviews, stacy makhnevich
Companies: medical justice

How Do You Promote A Bad DRM Scheme? With A Bad Movie Of Course

from the bad-gets-worse dept

Recently, we were talking about the horrible customer reviews that Warner's Ultraviolet DRM was getting. These customers were having such a hard time trying to justify the headache that came from complying with the requirements to watch their legally purchased movies. Warner seems to be trying to make the service worthwhile and is now bundling free digital movies with video games. If you buy Batman: Arkham City, you will also get a digital Batman movie for free. Which movie will you get? Batman Forever of course. Wait. What? If the goal of these promotions is to get people to start using the service, wouldn't you think they would try giving away a movie people would actually want to watch? As opposed to the worst Batman flick ever?

I guess the people running Warner just don't realize that customers want quality content. Sadly, when you are offering crap to customers there really isn't much you can do to polish it up. Maybe the next deal will actually be worth it. Most likely not though.

Filed Under: batman, batman forever, drm, reviews, ultraviolet, video games
Companies: warner bros.

Hollywood's Kinder, Gentler DRM: UltraViolet, Getting Slammed In Reviews

from the shocker dept

Remember UltraViolet? This was Hollywood's plan for a new, kinder, gentler DRM for movies that would try to provide some additional features. Or, rather, it would (oh so generously) grant you back just a few rights that anyone could get with an unauthorized version of a file... and you were supposed to thank them for this. As we noted at the time, what consumer problem does UltraViolet actually solve? The answer is absolutely none. It only attempts (and fails) to solve a perceived studio problem. Meaning that it was destined to piss off customers.

And that's exactly what's happening.

Jeff Malfant points us to an amusingly satirical Wired article, pretending to be from the studios' vantage point about UltraViolet:

UltraViolet will make purchasing a movie only slightly more of a pain in the ass than searching BitTorrent. Driving to a local video store and buying a copy on DVD or Blu-ray should only take about a half-hour more than downloading it, and I want to assure you that the legitimate copy will contain all the ads, auto-loading trailers and overproduced menu screens that even the pirates can’t figure out how to include.

Once you’ve bought it, all you’ll have to do is take the time to register it and give us whatever information we decide we need. We know pirates don’t have to do that, but we think you’ll find it fun. It’s sort of like Facebook, only instead of friends and family, you have a humongous powerful group of international corporations hanging on your every datum.

Once you’ve signed in, that’s when the fun begins. You see, the movies you paid for with your own money will be stored for you in your “locker.” Just like the lockers you use at school or the gym, they’ll be convenient, somewhat secure, they won’t actually belong to you and we can do anything we want with anything in them.

Let the pirates have their boring old “hard drives” and “networked media servers.” With UltraViolet, you’ve got a locker!

But even more telling, as sent in by Keith, are some of the user reviews popping up on Amazon that appear to totally slam UltraViolet as ridiculous and annoying:

This review does not relate to the movie, but it is focused on the ridiculous process to download the digital copy. After creating accounts for both Flixster and Ultraviolet, linking the accounts, enabling WB to view my personal information, the system hangs and doesn't download the movie. I contacted Ultraviolet first with the issues and error messages. After a day, I was told this is not an Ultraviolet issue, but a Flixster problem. I then contacted Flixster. They responded by sending me to the FAQ. To date, I have not gotten a proper response from Flixster on the error messages. I plan on canceling both accounts and will NEVER buy another DVD tied to Ultraviolet. This is a complete rip off and WB should be ashamed of this dreadful service. Please do yourself a favor and don't buy the movie with the digital copy. If you want a digital version, just got to iTunes.

And another:

First off I really liked the movie, if I was just rating the movie, it would probably be 3 1/2 stars, however, I have spent 4 days now trying to get this so called "digital copy". I knew what Ultraviolet was, and I thought it had promise to be used to stream to any device not locked to just one service (itunes), but it doesn't work. First off you have to sign into 4 different services to even get this stuff to maybe work, [...] some ultraviolet site and to "manage" your collection flixster collections...when in trying to link all of these useless accounts I can no longer access the movie to stream. So I am left with a product that does not work and customer service is a joke. I have been bounced around from site to site and from rep to rep who keep telling me to do the same thing over and over again, no matter how many times I tell them what has happened and what the issue is.

This flixster ultraviolet "suckloution" is the worst ever, you are better off ripping the DVD...

And another:

Warner Brothers has really pulled a fast one with this 3D edition that was supposed to include everything with this "UltraViolet" digital copy. It's NOT a digital copy, but rather an authorization to access a streaming video of the original cut of the movie online. This is NOT the package I pre-ordered! When I saw the commercials for the "UltraViolet" digital copy, I thought there would be some special feature-- I never dreamed that special feature would be WB promising me a digital copy that I couldn't use in iTunes EVEN THOUGH EVERY OTHER DIGITAL COPY OF A MOVIE DOES. The digital copy you do get (IF you can get the website to work!) has to be played on some kind of Adobe player THAT WILL NOT PLAY THE FILE. So you're stuck with their online streaming swill even though you were promised a digital copy.

And there are plenty more along those lines. When will the industry finally realize that DRM doesn't help anyone?

Filed Under: drm, reviews, ultraviolet

How Confusion & Lack Of Clarity In Copyright Law Make Reviewing Poems Difficult

from the can't-quote-this dept

techinabox alerts us to a short NY Times piece by David Orr, in which he notes yet another problem with copyright law, in that publishers are afraid to publish reviews of poems that actually dare to quote the poem, because it might be considered copyright infringement. At issue is that it's not clear what would be considered fair use in quoting a poem -- especially a short one. There are different theories about what's okay, but those theories don't seem to fit well with certain poems. Orr gives an example of a one line poem, and notes that quoting just 5% (a standard some suggest) would mean quoting 2 letters. As Orr notes the whole thing is an "incomprehensible system," and most publishers won't touch anything that might be interpreted as infringing, because they just don't want the headache of a legal battle.

Of course, quoting for the purpose of reviews should be fair use. The US Copyright Office itself uses:

"quotation of excerpts in a review or criticism for purposes of illustration or comment;"

as an example of what qualifies for fair use. But, tragically, the "excerpts" part of that standard is unfortunately vague, and it limits what people can do.

Filed Under: copyright, fair use, poems, poetry, reviews

Can A Computer Pick Out Fake Online Reviews When Humans Can't?

from the sounds-like-it dept

It's no surprise that there are a ton of "fake" reviews online of just about anything that can be reviewed. Businesses, hotels, authors, musicians, etc., all want to make sure that whatever it is they're selling, people see good reviews when they go searching. But, of course, that's a problem for consumers who rely on such fake reviews... and on the sites who host such reviews and want them to be as accurate as possible. So it's fascinating to see that some researchers at Cornell (yes, my alma mater) were able to come up with an algorithmic way to figure out what reviews are fake. You can read the full paper here (pdf). It's only 11 pages.

The method was pretty clever. First, they used Mechanical Turk to create 400 faked 5-star reviews of Chicago hotels:

To solicit gold-standard deceptive opinion spam using AMT, we create a pool of 400 Human- Intelligence Tasks (HITs) and allocate them evenly across our 20 chosen hotels. To ensure that opinions are written by unique authors, we allow only a single submission per Turker. We also restrict our task to Turkers who are located in the United States, and who maintain an approval rating of at least 90%. Turkers are allowed a maximum of 30 minutes to work on the HIT, and are paid one US dollar for an accepted submission.

Each HIT presents the Turker with the name and website of a hotel. The HIT instructions ask the Turker to assume that they work for the hotel’s marketing department, and to pretend that their boss wants them to write a fake review (as if they were a customer) to be posted on a travel review website; additionally, the review needs to sound realistic and portray the hotel in a positive light. A disclaimer indicates that any submission found to be of insufficient quality (e.g., written for the wrong hotel, unintelligible, unreasonably short, plagiarized, etc.) will be rejected

Then, of course, they need "real" reviews. But since part of the issue is that many "real" reviews are faked, the team did their best to find a bunch of real reviews from TripAdvisor, by narrowing them down based on a few factors:

For truthful opinions, we mine all 6,977 reviews from the 20 most popular Chicago hotels on TripAdvisor. From these we eliminate:

• 3,130 non-5-star reviews;
• 41 non-English reviews;13
• 75 reviews with fewer than 150 characters since, by construction, deceptive opinions are at least 150 characters long...
• 1,607 reviews written by first-time authors— new users who have not previously posted an opinion on TripAdvisor—since these opinions are more likely to contain opinion spam, which would reduce the integrity of our truthful review data...

Finally, we balance the number of truthful and deceptive opinions by selecting 400 of the remaining 2,124 truthful reviews, such that the document lengths of the selected truthful reviews are similarly distributed to those of the deceptive reviews. Work by Serrano et al. (2009) suggests that a log-normal distribution is appropriate for modeling document lengths. Thus, for each of the 20 chosen hotels, we select 20 truthful reviews from a log-normal (left-truncated at 150 characters) distribution fit to the lengths of the deceptive reviews.

They then test how humans see the two kinds of reviews, and discover that they can't tell them apart. In fact, their accuracy was only slightly above 50%. However, they then work out algorithmic ways of distinguishing the "real" reviews from the fake reviews, and come up with a system that is 90% accurate in picking out which reviews are which. Apparently, while humans can't pick out the differences, faked reviews have some common characteristics:

We observe that truthful opinions tend to include more sensorial and concrete language than deceptive opinions; in particular, truthful opinions are more specific about spatial configurations (e.g., small, bathroom, on, location). This finding is also supported by recent work by Vrij et al. (2009) suggesting that liars have considerable difficultly encoding spatial information into their lies. Accordingly, we observe an increased focus in deceptive opinions on aspects external to the hotel being reviewed (e.g., husband, business, vacation)...

[....]

... we find increased first person singular to be among the largest indicators of deception, which we speculate is due to our deceivers attempting to enhance the credibility of their reviews by emphasizing their own presence in the review.

Obviously, it's just one bit of research, but apparently those involved in it have been contacted by... well, just about everyone doing online reviews. Hopefully this means that we're not too far off from better quality online reviews.

Filed Under: fake, reviews