Even as companies around the world raced to comply with sweeping privacy rules that took effect in the European Union last month, EU lawmakers were working on another set of changes that could have a global impact on the internet.
Today a committee in the EU’s legislative branch approved a proposed model copyright law that would likely lead many apps and websites to screen uploaded content using automated filters to detect copyrighted material. The proposal will now move to a vote by the full European Parliament.
The effect would be similar to how YouTube tries to detect and block copyrighted audio and video from being posted on its site, but it would be applied to all types of content, including text, images, and software, as well as audio and video. Critics say this section of the proposal, Article 13, would lead to legitimate content, such as satire or short excerpts, being blocked even outside the EU.
Another section of the proposal would require online services to pay news publications for using their content. This has been widely referred to as a “link tax,” but hyperlinks and search engine listings are specifically exempted in the most recent draft of the directive shared by European Parliament member Julia Reda, a member of the Pirate Party Germany. The rules are widely seen as a way to force services like Facebook and Twitter that show short snippets or other previews of news stories to pay a fee to publishers, but the draft doesn’t make clear whether snippets would still be OK and, if so, how long they can be. The impact on Google is also unclear, as some of the material it displays, like its “featured snippet” information boxes, may not be considered search-engine listings.
The proposal is the latest attempt by European governments to reign in US technology giants. In addition to its privacy rules, the EU has in recent years imposed steep antitrust fines on Google, delivered Apple a hefty tax bill, and passed the digital “right to be forgotten.” Last year, Germany passed a law ordering social media companies to delete hate speech within 24 hours of it’s being published. Unlike these other rules, which focus on taxes and fees, the copyright proposal attempts to put more money into the pockets of publishers in Europe and elsewhere by mandating licensing fees.
A coalition of four European publishing groups released a statement applauding the European Parliament “for making a crucial stand for the future of a free, independent press, for the future of professional journalism, for the future of fact-checked content, for the future of a rich, diverse and open internet and, ultimately, for the future of a healthy democracy.”
The copyright proposal would be an EU “directive,” which would then be translated into laws in each EU country. Those laws could vary slightly. That, along with the vague wording of some parts of the proposal, make it hard to predict the exact outcomes of the rules.
Google head of global public policy Caroline Atkinson objected to the idea of preemptive filtering for all types of content in a 2016 blog post about an earlier version of the proposal. “This would effectively turn the internet into a place where everything uploaded to the web must be cleared by lawyers before it can find an audience,” she wrote. Atkinson wrote that paying to display snippets was not viable and would ultimately reduce the amount of traffic that Google sent publishers via Google News and search. Facebook and Twitter did not respond to requests for comment.
The proposal would shift the responsibility for publishing copyright-infringing work online from the users of a platform to the platforms themselves. It would mandate that services intended to store and publish copyrighted materials take “appropriate and proportionate measures” to ensure that copyrighted material is not available without the permission of its owner. It does not specify that sites must apply YouTube-style automated blocking, and it says that the “implementation of measures by service providers should not consist in a general monitoring obligation.” But critics argue that the directive will result in the widespread use of automatic filters. In some cases, platforms could avoid blocking content by licensing the content from rights holders.
The laws would only apply within EU countries, but companies might implement filtering around the world, says Gus Rossi, the director of global policy at the advocacy group Public Knowledge. He points to the way some companies, such as Microsoft, opted to follow the EU’s privacy rules globally, not just in Europe.
The way automated filters typically work is that rights holders upload their content to a platform like YouTube, and the platform’s software automatically watches for copies of those works. When the filter detects what it suspects to be infringing content, the platform blocks it from being published, or deletes it if it has already been published.
But critics say the filters will screen out content that should be legal, such as short excerpts from another work. In one ironic example, the French far-right political party National Rally (formerly known as the National Front), which supports the proposed copyright directive, recently had its YouTube channel briefly suspended due to alleged copyright violations, Techdirt reported. The channel is available again. National Rally did not respond to a request for comment.
Automated filters could be abused by people who don’t own the rights to content they try to protect, says Cory Doctorow, an author and special adviser to the Electronic Frontier Foundation. Someone could upload, say, the US Constitution to a site like Medium and claim it’s their copyrighted work. Then, if Medium had implemented an automated filtering system, the platform would block anyone from citing long passages from the Constitution. Doctorow says this could be abused by pranksters, or by those who want to suppress particular content. The draft proposal has no penalties for making false claims.
Automated filters could also be expensive for smaller companies to implement. “Far from only affecting large American Internet platforms (who can well afford the costs of compliance), the burden of Article 13 will fall most heavily on their competitors, including European startups” and small businesses, says an open letter signed by more than 70 internet pioneers, including web inventor Tim Berners-Lee and Wikipedia founder Jimmy Wales. The letter says filters will be unreliable, and the cost of installing them will be “expensive and burdensome.”
European Parliament member Axel Voss of the Christian Democratic Union of Germany admits that the proposal isn’t perfect and will likely lead to some false positives. But he tells WIRED that it will be better than the current system of allowing big platforms to profit by running advertising alongside copyright-infringing material. “We have to start somewhere,” he says.
Voss says the directive would affect a relatively small number of sites. Specifically, it would apply only to sites intended to be used to publish content and that “optimize” that content by doing things like categorizing it. The draft has exceptions for online retailers that mostly sell physical goods, for open-source software development platforms, and for noncommercial sites like online encyclopedias. But Reda argues that some sites might unintentionally be included under the rules, because the definition of which sites are covered is vague. For example, dating apps might have to screen the photos that users upload to ensure they don’t infringe copyrights.
The ultimate effect of the directive is murky, in part because it will be translated into law differently in different countries. That’s particularly problematic when it comes to defining when a site might need to pay to include a snippet or preview of a news article, since each country could come up with a different maximum amount of content that would be considered allowable.