Plagiarists, be warned. There’s a new browser tool called Churnalism that prowls the Web looking for blatant examples of cut-and-paste chicanery in news articles and other online content.

The Sunlight Foundation, a U.S.-based non-profit pursuing transparency in government and media, re-purposed a U.K. version of the tool. Churnalism functions as a website(Opens in a new window) where you can paste in URLs or text you’d like to test for plagiarism, but the Sunlight Foundation also built a browser add-on(Opens in a new window) that runs the software in the background on Internet Explorer, Firefox, and Chrome.

As a journalist myself, I was immediately excited and horrified by the existence of this tool. I was obviously going to test it on my own writing, hope for the best, and brace for the worst. We’ll get to that, but first some details about the nuts and bolts of Churnalism.

First, the bad. As a Web extension, Churnalism can be a little tricky to install. It loads easily enough into Firefox (it works with version 14 and higher) and Chrome (version 16+), but I had a heck of a time getting it to work with Explorer (versions 8 and 9). If you’re on 64-bit Windows 7, you have to install an additional add-on. Once I did get Churnalism installed, Explorer immediately began demanding that I disable it. After installing the add-on, I also got some flickering in different windows I had opened at the time, though that went away on a reboot.

The frequent prompts to disable add-ons that come with later versions of Explorer can be annoying but they’re actually very handy. They remind you of extensions you may have forgotten you installed and they let you know how much page-loading speed you’re compromising by having them—Churnalism is slowing me down by several seconds, according to Microsoft.

If you do install Churnalism, don’t fret if you don’t immediately see it working when you browse the Web. It turns out not all content on the Internet is plagiarized! When you come across a page with content the tool deems suspicious, an alert bar will appear at the top. You’ll have the option to dismiss the alert. Or you can click on the “Show Me” button for a pop-up window that lays out the suspicious copy next to the original source, with the questionable text highlighted in yellow.

When you’ve got Churnalism set up, you’ll be able to fiddle with the tool a bit. You can add websites you frequent to the tool’s “sites to examine,” which include familiar media sites run by the likes of The New York Times, The Economist, the Associated Press, and more. There are a few other ways to tinker with your Churnalism settings, including the choice to opt out of sending URLs you’ve visited to the foundation’s servers out of privacy concerns.

You can also use Churnalism as a website, where it’s possible to either paste in a URL or specific copy to determine if there’s some cut-and-paste hanky panky going on. Of course, Churnalism isn’t the first tool used to search for plagiarism online. In the old days, pasting a suspicious passage into Google’s search box and then sifting through the results also did the trick, if a lot less efficiently.

Churnalism runs on SuperFastMatch, which is an open-source search engine built by the Media Standards Trust, the developer of the original Churnalism site. SuperFastMatch checks copy against press releases from major corporations and the government, as well as Wikipedia. The detection process isn’t perfect—for example quotes or product names derived from a press release but properly attributed can still trigger an alert.

Continue Reading: Beware False Positives … and Negatives>

Beware False Positives … and Negatives

Beware False Positives … and Negatives
The Sunlight Foundation says it’s still working out the kinks and developing a “relevancy ranking that is derived from the total character overlap and the density of that overlap” to “filter out uninteresting matches and provide the best user experience.”

The takeaway is that you shouldn’t assume an article is plagiarized just because Churnalism flags it. You still have to compare the flagged article with the original source material to determine that the highlighted copy was truly lifted improperly instead of attributed in the proper manner.

Another issue is that it can take some time for Churnalism to catch up with questionable copy as it crawls through its resources looking for the original source. In the Churnalism tutorial video below, the Sunlight Foundation advises users to check back on really fresh Web copy that passes the plagiarism detector on the first go.

On to my own personal test of the Churnalism tool. After checking several stories I’d written in the past week or so, I was happy to learn that none of them set off Churnalism’s alarms via the browser extension. Just to be safe, I plugged those stories into the Churnalism website as well and still got a clean bill of health.

Then I tested this story from a couple weeks ago. The article in question is something I knocked out on the new Richland laptop chips from Advanced Micro Devices. To my chagrin, the copy was flagged by Churnalism(Opens in a new window).

The first suspicious passage highlighted by Churnalism was simply a product name, so no issue there. The second was an executive quote lifted from AMD’s press release, but correctly attributed. I was also happy to see that I’d specified that the quote was said “in a statement” by the executive—an unfortunately redundant turn of phrase but a useful one, because it tells the reader that the quote wasn’t directly spoken to the reporter but rather came from a press release.

The third section of copy Churnalism flagged was more problematic. It’s a bullet-point list of the various features built into the new Richland APUs and its lifted word for word from AMD. Now, if you read the sentence introducing that list, you’ll see that I specifically say that what follows is coming directly from AMD.

The problem is that there are no quotation marks around the list itself. That doesn’t invalidate the fact that there’s attribution of the original source, but it could be confusing to some readers. Why didn’t I use quotation marks? Simply put, they don’t work prettily with bullet-point lists in our system here at PCMag.

I didn’t do a thorough search of my article archive. But I’m pretty careful about that stuff—partly out of professional pride and ethics, partly out of fear of getting caught, and partly because as much as I’m frequently unsatisfied with my own writing, I’m vain enough to think it’s generally better than someone else’s.

The upshot is that Churnalism is going to make me take even greater care with my writing. Other journalists would be wise to do so as well. The Sunlight Foundation has produced a quick-and-easy tool for calling out plagiarists. I wouldn’t be surprised if some enterprising folks aren’t already deploying it on fishing expeditions through the archives of various online pubs just to see what they might turn up.