Category: algorithms

An Algorithm that Spots Fake Reviews and Exposes Sockpuppets

When I read about an algorithm that can now expose fake online reviews of any kind and expose sockpuppets, I found myself re-reading the article more than once. Evidently, someone’s invented new technology that will not only sort out fake reviews from real reviews, it can also expose sockpuppets and those sockpuppet names will turn hot pink all over the Internet so everyone knows they are socks. The fakes will be linked to their real identities forever. And, even more interesting, the blogs and web sites and social media of these sockpuppets and fake identities will also turn hot pink so everyone knows who they are.

Of course if you don’t use sockpuppets and fake identities online you have nothing to worry about. In fact, this new technology might turn out to be very entertaining. I can’t even imagine how the Internet will look if all sockpuppets and fake identities turned hot pink. Since I’ve never used a sockpuppet in my life, I’m not worried about it. I have posted anonymous comments at times for various reasons. I encourage anonymous comments here on this blog. And I do believe in freedom of speech, but I also believe that it’s human nature to take advantage of a good thing, so to speak, and you can only claim freedom of speech for so long with fake identities. As the old saying goes, “Give her in inch and she’ll take a yard.”

Now, if you believe what I just wrote above I have a bridge in Brooklyn for sale at a very reasonable price. No one’s going to turn hot pink any time soon. But it’s not a complete joke. I actually did read an article a couple of months ago about fake reviews, sockpuppets, and a new algorithm, but I haven’t had time to post anything about it. It’s an article written by, Josh Dzieza, a nice looking young guy who writes for a reputable publication from what I can tell. You can never be too sure anymore, so I always check that out first.

In the article Josh gets into the Jeffery Duns kerfuffle that happened not too long ago, when he (Duns) exposed JR Elroy for talking about his (Elroy’s) books on social media with fake identities.

  “This is RJ Ellory writing about his own book. And he has done this for them all, and yes, I’m proving it in the next few minutes,” Duns tweeted, before exposing Ellory’s pseudonyms.

Even though book reviews seem to get a lot of attention with regard to this topic, Josh isn’t just talking about book reviews. The article gets into all kinds of online corruption…it is what it is…and he mentions other web sites as well as Amazon. Although Amazon is probably the worst of the worst when it comes to reviews that NO ONE can believe…or should believe…it’s not just Amazon. (As a side note, I’ve almost decided that I’m not leaving any reviews on Amazon any longer. I’m not completely sure yet, but I’ve been holding off writing two reviews because I just can’t stand doing it anymore. It feels like a waste of time at this point. If I’m leaving honest reviews for books that I’ve read, it bothers me to know that those reviews that I took time to write are going to be up against fakes. I’ll post more about this in the future. I’m honestly still not sure how I feel yet and I might change my mind.)

But human eyes can go only so far. Fake reviews are ubiquitous on any site that lets users create anonymous accounts, such as Amazon, TripAdvisor, and Yelp; the tech research company Gartner projects that by 2014, between 10 percent and 15 percent of social-media reviews will be fake.

That’s interesting in itself, mainly to see that things like fake identities and sockpuppets are actually being studied and examined by people who seem to be taking it seriously. I’ve read more than a few blog posts since this past summer where some like Joe Konrath don’t seem to think it is serious. I’ve seen it joked about. And I have to wonder why some of these bloggers and writers don’t seem to think it’s serious. It’s obviously not just a few paranoid people talking, especially if companies like Gartner are studying the issue of fakes. We’re talking about major online fraud.

But this was what I found most interesting in Josh’s article:

 Since Duns unmasked Ellory, he has been bombarded with requests to investigate other suspicious accounts; he began looking into one of them, a famous author, and gave up. “There were thousands of reviews. You’d need an algorithm to sort through them.”

Such an algorithm is in the works. Last year Cornell researchers developed a program to detect suspicious hotel reviews on TripAdvisor. The researchers commissioned hundreds of fake hotel reviews using Amazon’s crowdsourcing site, Mechanical Turk, and isolated linguistic differences between genuine reviews and fake ones. They found that among other giveaways, fake reviews use the first-person frequently and pile on effusive adjectives and superlatives.

I would suggest reading Josh’s article in full. The most important thing I took from it wasn’t about whether or not sockpuppets and fake identities are good or bad, or whether or not freedom of speech comes into play. I’m not tying to sound “holier than thou” in this post and I’m not passing judgment on anyone. I’m just curious about one thing: if there is a problem/issue with online reviews, and if there is as much corruption as some claim there is, it sounds to me like there are going to be ways to reveal the fakes in the future and that’s not a chance I’d be willing to take with my name. So if you have done anything like this, you might want to think about the risk you’re taking and clean up your act a little. You just might wind up turning hot pink someday thanks to the same technology that gave the ability to use fake identities and sockpuppets in the first place.