Last week, Google expanded its Transparency Report to include data on notices it has received from copyright owners for links to infringing content. The Report now shows the number of notices, as well as their sources and targets, since July 2011.

The massive data release has triggered plenty of comment from news outlets and bloggers following copyright and internet issues. But there is one interesting fact drawn from Google’s Report that I want to highlight.

Copyright skeptics commonly try to make the claim that copyright enforcement is prone to abuse.1

These claims were especially prevalent during last winter’s debates over SOPA and PIPA. In an article on Popular Mechanics, for example, Adam Savage of Mythbusters said, “This is exactly what will happen with Protect IP and SOPA. We’ve seen it again and again. Give people a club like this and you can kiss the Internet as you know it goodbye.”2

The Google Report, however, paints quite a different story.

Google notes that, “From time to time, we may receive inaccurate or unjustified copyright removal requests for search results that clearly do not link to infringing content,” adding that it does not comply with such requests. How many of these requests are there? According to Google, “We removed 97% of search results specified in requests that we received between July and December 2011.” That means that out of all the requests Google receives, only 3% were sent by mistake or in bad faith.

Three percent.

Given the popularity of Google’s search engine3 and the large size of notices in the Report, it is reasonable to conclude that this percentage is representative of DMCA notices as a whole.

Google provides several examples of erroneous notices in its Report FAQ, and tech bloggers have had a field day reporting on these. But the fact remains that they represent a tiny sliver of the population of notices received. Among the noncomplying notices were also some made in apparent bad faith. These are never acceptable, but the number of them among total notices seems reasonable and does not indicate any systemic abuse. The level of “inaccurate or unjustified” requests reported by Google is in line with low-merit claims in other areas of the law — for example, a Harvard School of Public Health study found that 3% of medical malpractice lawsuits involved “no adverse outcomes from medical care.” And, it should be noted, the DMCA provides penalties for making misrepresentations on DMCA notices, penalties that have had negative consequences on several copyright owners — including some claims that copyright skeptics hold up as examples of copyright enforcement abuse.4

The Google Transparency Report reveals that the bogeyman of abusive copyright enforcement is nowhere near as frightening as copyright skeptics have repeatedly portrayed it to be. This is something that should be kept in mind the next time policymakers address the issue of protecting creator’s rights online.

Footnotes

  1. A few examples: The EFF has said, “We’ve seen that power abused time and again.” How PIPA and SOPA Violate White House Principles Supporting Free Speech and Innovation (Jan. 16, 2012); Rebecca MacKinnon and Ivan Sigal warned, “even existing copyright law is abused in attempts to stifle criticism and public debate” Online piracy laws must preserve Web freedom, CNN (Dec. 14, 2011); according to Mike Linksvayer of Creative Commons, “The DMCA was decried by advocates of free speech and the Internet, and has over past 13 years had many harmful effects.” Urgent: Stop [U.S.] American censorship of the Internet, (Nov. 11, 2011). []
  2. Other examples include: NetCoalition, a lobbyist group for Google, Facebook, and Yahoo, warned the bills have a “great potential for abuse by rights holders”; Rebecca MacKinnon of the New America Foundation stated, “Abuses under existing American law serve as troubling predictors for the kinds of abuse by private actors that the House bill would make possible.” Stop the Great Firewall of America, NY Times (Nov. 15, 2011); Julian Sanchez at TechDirt wrote, “This would be more convincing if the content industries weren’t so clearly continuing their long, proud tradition of making aggressive and overbroad copyright claims that would impede speech and innovation.” How SOPA Will be (Ab)Used (Dec. 19, 2011); Mike Loukides of O’Reilly Radar wrote, “there’s already a very lengthy history of copyright abuse by actors ranging from outright trolls such as Righthaven to supposedly reputable movie studios and record labels,” From SOPA to speech: Seven tech trends to monitor (Jan. 19, 2012); these and other claims can be summed up by the question asked by Alex Wexelblat on Copyfight: “It seems to me that the process is broken and people are not following the laws-as-written. How about we focus on fixing what’s on the books and in operation already before we go passing more new laws and further restrictions?” []
  3. According to Search Engine Watch, Google commanded two-thirds of the US market for internet search in February 2012. []
  4. For example, Adam Savage mentions a 2007 DMCA notice sent by Uri Gellar as an example of the kind of abuse that could “destroy the internet as we know it.” Gellar, however, ended up settling a subsequent lawsuit alleging misrepresentation for an undisclosed monetary amount and a court order to freely license the video footage that was the subject of the claim. []

15 Comments

  1. Michael Corliss

    That math is a little loose. Complying with 97% percent of requests does not mean that 97% of requests were legitimate; Google probably errs on the side of caution, because they lose safe harbor if they fail to comply with a valid request. Google could comply with 97% of requests even if 100% of the requests were sent by mistake or in bad faith.

    Furthermore, these results don’t include takedown requests for anything but search, and exclude requests for search that target Blogger or Youtube. There’s no reason to think that the pattern of take-down requests for search is the same as for a site like Youtube.

    • Re Michael Corliss:

      Your comment misses the point of Terry’s article. Terry is responding to those who looked at the exact same data and chose to focus on the 3% as proof that the DMCA is prone to abuse. It may well be that there are additional abuses occurring within the other 97%, but that’s not what is being argued by copyright skeptics and is besides the point. The fact is that they are using examples from the 3% to make their case. Terry’s point is to rightly take a big step back and to properly frame the discussion by showing that focusing on the 3% is misleading.

  2. 3% of how many complaints? Numbers are of little use without something to compare too, or work with.

    • The Report has those numbers. Google is currently receiving 1.2 million notices a month.

  3. I know we’re liable not to come to an agreement here, but I think your analysis is flawed. For one, as Michael points out above, the 3% figure is not a good representation of abuse, but a hard lower bound. Google has a lot more incentive to make over-comply than under-comply. I also agree with Michael that Search takedown numbers aren’t at all certain to resemble more content-y takedown numbers.

    But beyond that, I think there’s a deeper problem: those 3% are some 36,000 cases a month of speech that is being negligently or maliciously targeted, on just this one platform. Even if the percentage is constant, the absolute numbers obviously only grow across platforms. Tens of thousands of cases of a month, even among a million, seem to me to constitute a system “prone to abuse.”

    • How do you define “abuse” is the question? If studies show that 25% of online traffic is infringing, would you consider that a greater abuse of a system? How do we balance that?

      “Prone to abuse”? Careful, you’re kind of moving into prosecuting thought-crime here. People are “prone” to steal whatever isn’t bolted down; but we’re supposed to believe that filesharers are just taste-testing, right?

      Either way, I think some people here are mixing up confusion over the law with abuse. Glancing at the examples that Google provides paints a picture of people using DMCA notices in instances that aren’t even remotely copyright-related (most of the instances are concerns over alleged libel or companies sabotaging competitors). That’s abuse, i.e. corruption. However, just because a legitimate copyright holder might not understand the law surrounding appropriation of their work doesn’t mean they’re abusing the system. It just means that they’re confused, which is an argument often used to defend infringement.

      Concerning the numbers… so what? These are the numbers Google put out themselves. Mr. Hart doesn’t have a crystal ball. None of us know whether any of this info is accurate, honestly. Three percent seems pretty low to me. Hell, that’s less than what the big bad record labels pay out in royalties on the sale of a single album!

      those 3% are some 36,000 cases a month of speech that is being negligently or maliciously targeted

      Define “speech.” I fail to see how distribution of other people’s property, without their permission, is a form of speech. Is pawning stolen goods a form of speech, too? Isn’t plagiarism a form of speech? I guess I’m confused as to exactly when “speech” doesn’t mean “something spoken.”

      • Ah, nevermind about my last statement. I see what you were getting at: The people over whom Google was sent false DMCA notices. On the bright side, that’s 36,000 people (instances, really, multiple notices are sent for the same URLs) whose rights are being defended justly. That’s laudable.

        • Ah, nevermind about my last statement. I see what you were getting at: The people over whom Google was sent false DMCA notices. On the bright side, that’s 36,000 people (instances, really, multiple notices are sent for the same URLs) whose rights are being defended justly. That’s laudable.

          Assuming 3% of notices are in bad faith, that’s 36,000 instances of wrongdoing that in all likelihood will go unpunished. It shouldn’t be that way, and I think that those people should be punished. But that’s nothing. Assuming the other 97% of notices are in good faith, then that’s 1,164,000 instances of wrongdoing that will also go unpunished. It shouldn’t be that way either–those people should be punished, but instead nothing happens to them. I don’t see Parker complaining about these millions of intentional wrongdoers who violate other people’s rights and get away with it. Funny that.

          • 36,000 instances of wrongful infringment a month are nothing? So 432,000 censorship notices in a year mean nothing to you? How about 4,320,000 instances of censorship in a decade?

            And yet, putting the burden on the people to prove their innocence through this system that is already heavily favored towards the RIAA must be for the best. And who has proven all of the wrongdoing in a court of law? Seeing the report, the RIAA sure has more instances of “friendly fire” than instances of wrongful service.

            I never thought that someone could honestly advocate for moral censorship that hasn’t done a thing to stop people from finding what they want. Sad to see that people still can’t figure out that the transparency report shows just how fruitless this fight against piracy and websites really is. Not only does the taking down of content not prevent piracy, but looking at the websites, people still go to the ones not authorized by the studios. I guess it’s time for plan ZZ now, eh?

    • I know we’re liable not to come to an agreement here, but I think your analysis is flawed. For one, as Michael points out above, the 3% figure is not a good representation of abuse, but a hard lower bound. Google has a lot more incentive to make over-comply than under-comply. I also agree with Michael that Search takedown numbers aren’t at all certain to resemble more content-y takedown numbers.

      But beyond that, I think there’s a deeper problem: those 3% are some 36,000 cases a month of speech that is being negligently or maliciously targeted, on just this one platform. Even if the percentage is constant, the absolute numbers obviously only grow across platforms. Tens of thousands of cases of a month, even among a million, seem to me to constitute a system “prone to abuse.”

      The system that’s being abused is the copyright system. And it’s being abused on a massive scale by the pirates–they are causing this. Yet your side of the debate seems to have no problem with that abuse. I don’t understand the obsession with putting every single thing a rights holder does under a microscope, but bad faith pirates who willfully violate other people’s rights get a free walk. It’s ridiculous. If someone files a takedown notice in bad faith, then they should answer for their misrepresentations. But a pirate who simply has his infringing file taken down or removed from a search result is actually getting away with something too. And I don’t see you calling for them to answer for their wrongs. The bias is palpable.

  4. Those who have blindly fixated on insisting on stopping anyone, anywhere from being able to copy anything, without ever once seriously considering the uniform sea of authoritative voices who are screaming that to do so with destroy the Internet, their catastrophically evil selfishness cannot possibly be overestimated.

  5. Pingback: Finger on the Pulse: From Our Blogroll and Beyond « The Legal Pulse