While the directive fixes some longstanding problems with EU rules, it creates much, much larger ones.
By Cory Doctorow / 06.07.2018
The pending update to the EU Copyright Directive is coming up for a committee vote on June 20 or 21 and a parliamentary vote either in early July or late September. While the directive fixes some longstanding problems with EU rules, it creates much, much larger ones: problems so big that they threaten to wreck the Internet itself.
Under Article 13 of the proposal, sites that allow users to post text, sounds, code, still or moving images, or other copyrighted works for public consumption will have to filter all their users’ submissions against a database of copyrighted works. Sites will have to pay to license the technology to match submissions to the database, and to identify near matches as well as exact ones. Sites will be required to have a process to allow rightsholders to update this list with more copyrighted works.
Even under the best of circumstances, this presents huge problems. Algorithms that do content-matching are frankly terrible at it. The Made-in-the-USA version of this is YouTube’s Content ID system, which improperly flags legitimate works all the time, but still gets flack from entertainment companies for not doing more.
There are lots of legitimate reasons for Internet users to upload copyrighted works. You might upload a clip from a nightclub (or a protest, or a technical presentation) that includes some copyrighted music in the background. Or you might just be wearing a t-shirt with your favorite album cover in your Tinder profile. You might upload the cover of a book you’re selling on an online auction site, or you might want to post a photo of your sitting room in the rental listing for your flat, including the posters on the wall and the picture on the TV.
Wikipedians have even more specialised reasons to upload material: pictures of celebrities, photos taken at newsworthy events, and so on.
But the bots that Article 13 mandates will not be perfect. In fact, by design, they will be wildly imperfect.
Article 13 punishes any site that fails to block copyright infringement, but it won’t punish people who abuse the system. There are no penalties for falsely claiming copyright over someone else’s work, which means that someone could upload all of Wikipedia to a filter system (for instance, one of the many sites that incorporate Wikpedia’s content into their own databases) and then claim ownership over it on Twitter, Facebook and WordPress, and everyone else would be prevented from quoting Wikipedia on any of those services until they sorted out the false claims. It will be a lot easier to make these false claims that it will be to figure out which of the hundreds of millions of copyrighted claims are real and which ones are pranks or hoaxes or censorship attempts.
Article 13 also leaves you out in the cold when your own work is censored thanks to a malfunctioning copyright bot. Your only option when you get censored is to raise an objection with the platform and hope they see it your way—but if they fail to give real consideration to your petition, you have to go to court to plead your case.
Article 13 gets Wikipedia coming and going: not only does it create opportunities for unscrupulous or incompetent people to block the sharing of Wikipedia’s content beyond its bounds, it could also require Wikipedia to filter submissions to the encyclopedia and its surrounding projects, like Wikimedia Commons. The drafters of Article 13 have tried to carve Wikipedia out of the rule, but thanks to sloppy drafting, they have failed: the exemption is limited to “noncommercial activity”. Every file on Wikipedia is licensed for commercial use.
Then there’s the websites that Wikipedia relies on as references. The fragility and impermanence of links is already a serious problem for Wikipedia’s crucial footnotes, but after Article 13 becomes law, any information hosted in the EU might disappear—and links to US mirrors might become infringing—at any moment thanks to an overzealous copyright bot. For these reasons and many more, the Wikimedia Foundation has taken a public position condemning Article 13.
Speaking of references: the problems with the new copyright proposal don’t stop there. Under Article 11, each member state will get to create a new copyright in news. If it passes, in order to link to a news website, you will either have to do so in a way that satisfies the limitations and exceptions of all 28 laws, or you will have to get a license. This is fundamentally incompatible with any sort of wiki (obviously), much less Wikipedia.
It also means that the websites that Wikipedia relies on for its reference links may face licensing hurdles that would limit their ability to cite their own sources. In particular, news sites may seek to withhold linking licenses from critics who want to quote from them in order to analyze, correct and critique their articles, making it much harder for anyone else to figure out where the positions are in debates, especially years after the fact. This may not matter to people who only pay attention to news in the moment, but it’s a blow to projects that seek to present and preserve long-term records of noteworthy controversies. And since every member state will get to make its own rules for quotation and linking, Wikipedia posts will have to satisfy a patchwork of contradictory rules, some of which are already so severe that they’d ban any items in a “Further Reading” list unless the article directly referenced or criticized them.
The controversial measures in the new directive have been tried before. For example, link taxes were tried in Spain and Germany and they failed, and publishers don’t want them. Indeed, the only country to embrace this idea as workable is China, where mandatory copyright enforcement bots have become part of the national toolkit for controlling public discourse.
Articles 13 and 11 are poorly thought through, poorly drafted, unworkable—and dangerous. The collateral damage they will impose on every realm of public life can’t be overstated. The Internet, after all, is inextricably bound up in the daily lives of hundreds of millions of Europeans and an entire constellation of sites and services will be adversely affected by Article 13. Europe can’t afford to place education, employment, family life, creativity, entertainment, business, protest, politics, and a thousand other activities at the mercy of unaccountable algorithmic filters. If you’re a European concerned about these proposals, here’s a tool for contacting your MEP.