Google Anti-Piracy Agreement Will Target Domain Hopping, Share Search Data

More details surrounding the voluntary anti-piracy agreement between Google, Bing, and major rightsholders have emerged. In addition to efforts to mitigate pirate sites' efforts to jump domains to avoid downranking, the search engines will "exchange detailed information" with rightsholders "on a confidential basis" in order to better understand how users are searching for content.

Following roundtable discussions between the BPI, Motion Picture Association, Alliance for IP, plus representatives of Google and Bing, in February a voluntary anti-piracy agreement was announced.

Under this anti-piracy code, search engines agreed to further optimize their algorithms to demote pirated content in search results, with the aim of making infringing content less visible and legal alternatives easier to find.

As highlighted last month, details of the arrangement were planned to remain largely secret but thanks to a pair of Freedom of Information (FOIA) requests from both the EFF and TJ McIntyre from Digital Rights Ireland, we now have a somewhat clearer idea of what will be happening between the groups.

As expected, the main focus is the search deranking of sites “dedicated to infringement” based on the volume of valid DMCA-style notices rightsholders send to Google. In other words, sites that index a majority of infringing content and are subjected to a lot of rightsholder complaints will find their results buried.

Unfortunately, the report released under the FOIA request is redacted, which leaves some sections ambiguous at best and hard to follow at worst. This section, however, seems to indicate an additional effort to manipulate search results that are generated from “neutral” non-piracy related search terms.

“All parties support the objective of removing links to infringing content from [REDACTED] search results returned to consumers in the UK in response to ‘neutral’ formulations of search query (exact search terms to be agreed) with the goal of presenting the consumer with links to legitimate sites. This includes search results presented to the user in the form of natural search results, sponsored or advertisement results or media player ‘box’ results,” the document reads.

These search terms haven’t been defined publicly but based on previous copyright holder complaints, words such as ‘download’, ‘MP3’ and even artist or content names could be in the mix. In any event, an assessment will take place to see how they prejudice rightsholders, especially when it comes to fresh content.

“Selection of such search queries by the parties shall take into account data indicating the actual levels of usage of such search terms, as well as the harm that illegal access to content via specific queries can cause to creators, in particular for new releases,” the agreement notes.

Google’s AutoComplete feature, which has proven controversial in the past, will also be subject to tweaks that focus on not suggesting infringing content when neutral terms are entered.

Moving forward, an area that is likely to raise an eyebrow or two is a statement in the agreement which possibly suggests the sharing of search engine user behavior data with rightsholders.

“Search engines and rights holders will exchange detailed information on a confidential basis in order to better understand how users are searching for content,” it reads.

“This information exchange will not be expected to include commercially confidential information, and is without prejudice to the existing legal remedies available to either party.”

To give an indication of how complex these discussions must’ve been at times, one only has to look at the following paragraph, which appears to be an effort to lay some of the blame with rightsholders, should infringing links appear more prominently than legal ones in search results.

“Performance in achieving the above metric should be considered in tandem with an objective assessment of the existence of legitimate websites (of rights holders or their partners, distributors or other authorized locations) that offer consumers access to legitimate content or information for the measured queries, and the efforts made by rights holders to take advantage of reasonable techniques such as search engine optimisation,” it reads.

In other words, rightsholders shouldn’t be able to blame Google and Bing for the appearance of ‘pirate’ results if they don’t make legal alternatives available or fail to carry out effective SEO. That shouldn’t be too much of a problem though, since the agreement notes that the parties will work together to optimize SEO for legitimate sites to “improve the likelihood such sites will rank higher in results for well-meaning queries.”

Another interesting detail in the agreement is how the parties intend to tackle so-called “domain hopping” by pirate sites. Currently, when Google receives a lot of DMCA notices for a domain utilized by a pirate site, the site is downranked in results. That often leads to the site getting a new domain, at which point the ‘clean’ domain starts appearing higher in results again. The agreement seeks to deal with that.

“All parties will work with the [Intellectual Property Office] to evaluate how frequently copyright infringing websites, subjected to demotion, change their top-level domain (TLD), but otherwise retain substantially the same identity,” the agreement reads.

“If this activity is sufficiently widespread as to justify it, search engines and rights holders should develop a process whereby rights holders can notify search engines of the occurrence so that, when verified, such domains can be appropriately demoted.”

Overall, Google and Bing will work with rightsholders to demote domains quicker, with the latter encouraged to use APIs and better-formatted infringement notices. A whitelist of sorts will also be introduced, to ensure that legitimate sites don’t get caught up in Google and Bing’s downgrading filters.

But for those concerned about the potential for this voluntary agreement to spread beyond those currently involved, there’s something looming on the horizon. Google and Bing have also committed to sharing their work in this area with search engines and rightsholders that are not already signatories.

“All parties to this Code of Practice commit to ensure that progress or best practice in this area (to the extent that such information is non-confidential) is shared widely with smaller search engines and independent rights holders,” the agreement notes.

As previously reported, the Minister of State for Intellectual Property will oversee the implementation of the voluntary code, and provide quarterly cycles of research and a review after one year.