4 Comments

Summary:

Google today released a new picture of the millions of links it scrubs from its search results in response to requests from Microsoft, movie studios and other content owners.

Google today released a new picture of the millions of links it scrubs from its search results in response to requests from Microsoft, movie studios and other content owners. In a reflection of the evolving nature of anti-piracy enforcement, the company revealed that it takes down 250,000 search links each week over copyright concerns, a figure that exceeds the total number it removed in all of 2009.

The data arrived today as a new section in Google’s Transparency Report, a set of findings that show how governments — and now private actors — are removing pages from the internet.

Google’s senior copyright counsel, Fred von Lohmann, stressed in an interview that the vast majority of the takedown requests are legitimate and come in response to sites offering unauthorized copies of software, entertainment or pornography.

Here is a screenshot from the Transparency Report that shows who is ordering the take downs and which websites most commonly host  unauthorized content:

The screenshot also shows how the number of copyright requests is growing exponentially. Google did not say whether this spike is the result of an increase in piracy or instead is due to more sophisticated tools that make it easier for rights owners to detect when their content has been misappropriated. Google says it’s processing requests faster than ever and that its average response time to a takedown demand is now 11 hours.

The figures are sure to add grist to the debate about the prevalence and the reasons for online piracy. Content producers have long complained that internet companies don’t do enough to remove infringing material. On the other hand, journalists and civil libertarians have argued that copyright owners have been too aggressive in their enforcement tactics.

In a blog post on the findings, Google cites an example in which “a major entertainment company” demanded the removal of a search result that linked to a review of a TV show despite the fact no copyright infringement had occurred. When sites are removed, Google places a notice in its search results and also forwards the information to Chilling Effects, a website run by the EFF and major universities that reports on activities that chill free speech.

But von Lohman added that requests like the one by the entertainment company are often simple errors rather than an attempt to repress criticism. He noted that Google complies with 97 percent of all requests using a combination of human and algorithmic reviews, and that the data is a validation of the current DMCA copyright regime in which internet companies are not liable for the actions of bad actors so long as they comply with notices by rights owners.

The new copyright elements of the Transparency Report may also help Google and other technology companies head off fresh attempts by Hollywood to pass legislation like the failed Stop Online Piracy Act (SOPA).

von Lohmann didn’t elaborate on Google’s strategic reasons for releasing the information but did say it will allow policy makers to “consider the data” when evaluating copyright law.

Update: Techdirt reports that Microsoft, which leads the list in takedown requests, did not remove many of the same offending links in its own Bing search engine.

  1. CJ Millisock Thursday, May 24, 2012

    Does 1.2m sound EXTREMELY low to anyone else?

    Share
    1. Q3 technologies Friday, May 25, 2012

      @CJ I thought the exact same thing. And after getting it down to a micro level it comes down to about 27 requests/min. Which is relatively low considering the amount of users on Google. But the alarming bit is that this number is increasing exponentially.

      Share
  2. this article has a good example of innumeracy in a report of numerical data:

    “The screenshot also shows how the number of copyright requests is growing exponentially.”

    Curve fitting the data from Google’s site shows that an exponential fit is very poor (R squared ~ 0.67) and worse than a linear fit.

    However alarming it may be that Google’s takedown requests have roughly doubled or tripled in less than a year, they haven’t grown exponentially, at least according to the data.

    Share
    1. they are bloggers! hype&viralness > correctness

      Share

Comments have been disabled for this post