The EU's New Copyright Law Changes the Internet's Free-Content Principle — and Could Bring 8 Unsettling Futures

For the long-running conflict between internet platforms and content creators, this may not be the end of one problem but the beginning of many.

The global tech giants' natural enemy has struck again.

After two and a half years of protests, debates, petitions, and repeated votes, the European Parliament last week (March 26) passed the Directive on Copyright in the Digital Single Market — the new copyright law — by 348 votes to 274.

Julia Reda, a member of the European Parliament, then posted the result for protesters on Twitter. She called it "a dark day in the history of internet freedom." By Reda's count, 5 million people have signed a petition against the reform so far.

Google and other tech giants were as disappointed as Reda. "Google News may withdraw from the European continent because of this directive," said Jennifer Bernal, Google's public policy manager for Europe. Over the past year the company had lobbied the European Parliament repeatedly, urging it to amend the bill.

In recent years the EU has become known for the size of the fines it hands American tech companies and the severity of the laws it writes for them. Fake news, privacy, ad transparency — every fine and every bill has come with an argument, but usually with supporters inside the tech industry too. Yelp has accused Google of unfair competition; Spotify has just filed an antitrust complaint against Apple's App Store. A law drawing near-unanimous opposition from the whole industry, as this one has, is new.

The opposition includes giants like Google and Facebook; consistently neutral open platforms like GitHub and Wikipedia; people and organizations that are normally wary of what big tech does, such as the Electronic Frontier Foundation (EFF) and World Wide Web inventor Tim Berners-Lee; and even the governments of 11 EU member states.

The supporters are easy to picture: musicians, writers, and screenwriters outside the Parliament in Strasbourg, France, holding up yellow signs reading "#yes2copyright," blowing trumpets and beating drums to celebrate the vote. That camp includes more than 270 organizations representing writers, musicians, directors, and producers, among them the International Association of Scientific, Technical and Medical Publishers (STM), the European Grouping of Societies of Authors and Composers (GESAC), and the European Newspaper Publishers' Association (ENPA).

In protest, Wikipedia had earlier taken its German, Czech, Danish, and Slovak pages offline for a day. The protest worked: the final text of the bill exempts Wikipedia by name.

Two articles caused most of the argument.

Article 15 says online platforms may not use a creator's content without permission, even a fragment of it — an article's headline and summary, for instance. News outlets also gain the right to charge third-party aggregators (Google News and the like) that display their reporting. The charge has been nicknamed the "link tax."

Article 17 says platforms that run on user-generated content (UGC), such as Facebook and YouTube, must block unauthorized copyrighted material at the moment of upload. This screening is known as the "upload filter." Until now, platforms could let content go up first and take it down after a rights holder complained.

EU law governs an enormous market — more people than the United States, more spending power per head than China.

Its effects will not stop at Europe's borders. EU law sets an example: the GDPR, its earlier privacy law, has already prompted talk of copying it in several countries. Just last weekend, Mark Zuckerberg wrote that the U.S., and governments everywhere, should follow the EU and pass GDPR-like rules that draw clear lines around tech companies.

The EU Is Trying to Solve a Real Problem

Most supporters of the law make the same argument: platforms built by companies like Google earn enormous advertising revenue from copyrighted content and pass only a sliver of it back to the people who made that content. This is unfair.

Google and Facebook have long insisted that they share revenue fairly with rights holders, but the reality is less tidy. Take Google Search. A website can run Google ad banners and take a cut of the traffic once a user clicks through. Most of Google's revenue, though, is made before that click — on the results page itself.

The same is true of the ads on Facebook, WeChat Moments, Baidu's homepage, the new Google homepage, and Toutiao: they sit inside the feed. That is where the money comes from, and none of it has to be shared with whoever supplied the content.

Google argues that its search platform sends the news industry more readers, which sounds like an obvious benefit. But an EU investigation found that fewer than 50% of readers click the news links on Google's search pages. Most read the headline and the summary and stop there — by which point Google has already sold the ads in the feed and down the side of the page.

In 2001, Google's advertising revenue was $70 million. In 2018 it was $116.32 billion — but the share coming from ads placed on other people's pages, the ad network business, has fallen every year.

In the ad network's first years it brought in almost half of Google's advertising revenue. By 2018 it was 17%. The slice Google has to split with websites has shrunk from nearly a half to less than a fifth.

Facebook has gone further than Google: most of its ad slots sit inside the feed. In May 2015 it launched Instant Articles, meant to help third-party partners publish news inside the social network while selling ad space in their own articles. But in January 2018 the Tow Center, analyzing links on Facebook, found that more than half of Instant Articles partners had abandoned the format.

In 2018 Facebook's advertising revenue reached $55.013 billion, 92% of it from mobile ads.

Revenue keeps rising at the companies that distribute content and keeps falling at the companies that make it — above all in a news industry that has moved online. In 2003, U.S. print and online news outlets together took in $46 billion in advertising. By 2014 the whole industry's ad revenue was $20 billion. By 2016 the number of American journalists had fallen a full 40% from the 456,000 counted in 1990.

News sites are not the only casualties. In March 2011, comScore research showed that the time users spent on Facebook had grown 69% over the previous 12 months while time spent on independent sites outside Facebook had fallen by nearly 9%. Before Facebook's mobile app had even taken shape, the old Web was already collapsing and a new internet was rising in its place.

Netcraft counts the world's websites once a month. Its data shows the fast-growth years of the early 2000s are over: the number of active sites worldwide has been flat since 2011, hovering around 182 million and slowly declining. In the first three months of 2019 alone, hundreds of thousands of old websites went dark.

There are fewer people supplying content to the internet. Over the same eight years, the tech giants that dominate social media and search have multiplied their advertising revenue more than a thousandfold.

The problem is real and the conflict keeps growing. The question is whether the EU's answer solves it.

After Breaking the Internet's Tradition of Free Content, the New Copyright Law Could Bring 8 Unsettling Futures

Faced with a real problem, the EU's answer is to pull content that has been free by custom since the internet began under copyright.

Google started it: search engines scan whole websites and huge numbers of books, then match everything against the keywords you type. All of it rests on the "fair use" principle in the Digital Millennium Copyright Act.

That law guarantees that as long as Google stays inside the framework it bears no liability. It can gather as much as it can and take material down promptly when a problem arises.

That has been the foundation of the internet's growth for years.

The bill now has the European Parliament's approval, but the EU still has not said how copyright fees will actually be paid. Nobody knows how it will work in practice. Will tech companies have to hand over copyright revenue on their own initiative, whether or not the rights holder asks for it, or will rights holders have to approach the companies and claim it? Nor is it clear how the fees will be calculated. Different arrangements would send the story in very different directions.

Even so, charging a copyright fee for something like an article summary is by itself enough to change how information moves around the internet. Whichever arrangement wins out, none of the likely outcomes looks good.

1. Fewer search results, and small rights holders pushed out of the market

If aggregators really must pay every rights holder before showing their work, the amount we can find through Google Search will shrink sharply.

Google's free search service was built on a copyright law that let it index other websites' full text for nothing. That permission was later narrowed to links, headlines, and summaries. Under the new law the free-indexing window narrows again, to "a few words" or "a single sentence" — and at that point platforms have every reason to index less, not more.

Richard Gingras, Google's vice president for news, wrote that if the law takes effect Google may end up dealing with only a handful of large rights holders, buying books, news, and music from them — which would leave most unknown work with no chance of being seen.

In Gingras's words, that puts Google in the position of deciding who gets exposure and who does not, who wins and who loses. The incentive is real: if a summary costs money, a company will buy the content with the largest audience to get the most for its spending.

The German research outfit Golem estimated that if every piece of copyrighted content had to be bought before it could be shown, small publishers would take less than 1% of the copyright money and the biggest publishing groups would take 64%.

At the extreme, work by individuals and small publishers would get no exposure at all from big aggregators like Google, and would die out on its own. Our culture and our public argument would be left with mainstream voices only.

2. Worse search results, and creators pushed out of the market

This is the opposite scenario, and which one arrives depends on how each internet company chooses to comply.

To protect the quality of its results, Google might do exactly what Richard Gingras described: spend heavily on content from a few top rights holders, an expense with no end to it.

But Google also has to control costs — and not every aggregator earns Google's tens of billions of dollars in annual profit.

Small news sites are nothing like The New York Times or The Wall Street Journal, with their thousands of editorial staff. They do not need many reporters or editors and do not carry the daily cost of reporting and investigation. Turning out news in bulk — even fake news — costs them very little.

In exchange for exposure, sites that produce cheaply will often accept a lower copyright fee, or none at all.

Economists call this the lemons problem: the market's mechanism for keeping the good and weeding out the bad gets bent backwards, so good products are pushed out and bad ones stay, and at the extreme the market shrinks until it disappears. Buyers will only pay a price based on the average quality they can perceive, which is low enough to suit sellers of poor goods, while sellers of good ones, unable to get a price that matches their quality, leave.

3. Access to information online, once relatively equal, becomes unequal

At a time when society as a whole is growing more unequal, the internet has been the exception.

Information equity is a term born with the internet. It means fairness in how information is spread among individuals, groups, and regions — the idea that everyone should have a chance to reach the information that matters to their lives.

Sergey Brin, one of Google's founders, declared at the company's founding that Google's goal was "to organize and understand all the information in the world with a perfect search engine." Over the past twenty years Google's existence has helped information travel freely on the internet and pushed information equity forward.

Free is the best way to spread information. Once platforms must split revenue with rights holders before any revenue exists, the degree of "free" is bound to change.

In today's United States, where the paid model rules, outlets large and small have built paywalls. People who will not or cannot pay tens of dollars a month in subscriptions can barely reach primary information at all, and only by paying still more can you read across a range of viewpoints. Information equity is already eroding — and that is before the new copyright law.

Jennifer Bernal did raise the possibility, after the vote, of Google's news business leaving Europe. That too has been read as an empty threat.

Francois Godard, a European media analyst at the research firm Enders Analysis, said: "I don't believe that threat. Google really needs Europe."

4. Higher barriers to entry, and a deeper tech-giant monopoly

Small aggregators may be the group the new law hurts most. Articles 15 and 17 require them, respectively, to buy copyrighted content and to build a copyright-filtering system. Both cost a great deal.

On the ability to buy content, a small company limited by its assets can never outbid Google.

And if it chooses to show only free content, the quality will never match Google's either. It is like restricting an image search to pictures licensed for third-party use: the results will never be as good as the unfiltered ones.

"No copyright restrictions" tends to mean "poor quality." That has held true in every corner of the internet.

Building a copyright filter is a major expense in itself. YouTube is the reference point: it spent at least $60 million building the automated system that handles copyright claims.

The European Parliament thought about this, and the final text exempts small companies:

Small platforms will benefit from a lighter regime where they have not obtained authorization from rights holders. Online service providers that have operated in the EU for less than three years, with turnover below 10 million euros and fewer than 5 million monthly users, need only show that they have made "best efforts" to obtain authorization and have acted quickly to remove unauthorized work flagged by rights holders. Once such a company's monthly unique users pass 5 million, it must also show it has made best efforts to keep flagged work from reappearing on the platform later.

But "showing best efforts" is a very loose standard. And even if a small company can post first, take down on request, and never repost, the content it is allowed to show will always be capped by how much money it has.

When regulation tightens, the cost of doing business rises and newcomers find it harder to challenge the incumbents. Banking, oil, insurance — the pattern holds everywhere.

5. Some works will disappear altogether

Google went through its own wave of copyright trouble around 2005. It had just begun building an online library, scanning books on a large scale with publishers' cooperation. The plan then was to work with the big publishers and share the revenue with them.

It still landed Google in court. In 2006, Wired senior writer Kevin Kelly set out the heart of the problem in The New York Times: of everything humanity has created, the books whose copyright sits with a major publisher are a small fraction. Seventy-five percent of books are in fact stranded on "copyright islands."

This was the result of the "new copyright law" the U.S. Congress passed in 1976. Under it, creators no longer had to register or renew copyright with the government; the simple act of making a work granted its maker copyright. For artists that looked like good news and an encouragement to create. What followed was that more and more copyrights belonged to individuals rather than companies, and much of that work was never taken up by a publisher and slowly sank into the dark.

One by one, those "copyright islands" merged into a "copyright continent." For Google, finding the owners of that 75% of books and negotiating a license with each of them one at a time was all but impossible.

"No one thinks all of that 75% of unpublished work is a masterpiece," Kevin Kelly wrote, "but for the world of knowledge, if we don't use technology to record them, huge numbers of books humanity has created could vanish completely within a few decades."

In 2004 Google proposed a solution: for books without a copyright clearance it would show readers only a few selected sentences, something like an academic abstract. Google's lawyers argued this should count as "fair use."

That is how Google's online library came to look the way it does now — the cover if one exists, the author, a description, excerpts, a table of contents.

The lawsuits kept coming anyway. The big publishing houses that had never published that 75% of books argued Google was depriving them of a chance to profit from it.

So Google offered another solution: for works whose ownership could not be established, upload first and delete later. It argued that it had to scan and upload these books first, and would then honor any legitimate request to take the scans down.

More and more internet platforms copied this do-it-first-and-answer-for-it-later approach. It is also where "take it down if this infringes" — the two-word disclaimer that Baidu Tieba and Zhihu users attach to reposts — comes from.

The EU's new law closes off that route. First, fair use shrinks to "a single word" or "a very short sentence"; second, tech companies now have to be sure there is no copyright problem before anything goes up rather than after. Both legs the online library stood on have been kicked.

Since the bill was first published, Google's ask has been the same: widen fair use. Not "a single word" or "a sentence," but summaries and thumbnails — "basic information that lets users understand the content."

The final text did not give it that. If online libraries are rebuilt strictly to the letter of the law, books with murky copyright or little copyright value really will be included less often, or disappear from the internet entirely.

For young people today, what cannot be found online does not exist.

6. Search engines and social networks will send websites less traffic, concentrated in fewer places

On a search page, in WeChat Moments, on a news aggregator, the images, the summary, and the full headline are what make people click. Internet companies worked that out over years of practice.

Kent Walker, Google's senior vice president, wrote that Google ran an experiment: rather than pay outlets a licensing fee, it did what the bill demands and showed only part of a headline plus the bare link, with no information about the story behind it. Traffic to publisher sites from Google News fell 45%.

Richard Gingras wrote that Google sends news sites more than 10 billion clicks a month, and cited a Deloitte study putting the average value of a user visit at 0.04 to 0.08 euros. "Thousands of news publishers use Google's advertising services and keep 70% or more of the revenue."

Groups representing publishers said in a joint statement that they saw the experiment as a stunt. Google, they said, would rather offer a worse product than pay for the work of journalists and publishers. "If Google really does this, people will abandon Google."

7. Copyright trolls multiply, and derivative work is squeezed

The new law requires tech companies to run "upload filters" that stop users from posting material that infringes someone else's copyright. That has alarmed the people who make memes, sticker packs, film and TV edits, and game commentary.

Craig Thompson, a game-commentary YouTuber with nearly 500,000 subscribers, said of the law: "Gamers are done. You're done. I'm done."

The European Parliament considered this too, and addressed it in the bill:

Q: Will the Copyright Directive stop users from expressing themselves the way they do now? Will memes and GIFs be banned? A: No. The EU specifically allows users to upload memes and other content made for quotation, criticism, or parody. Users will be able to go on creating this kind of content online.

Reality may be less simple.

In 2012 a healthy-living YouTuber named Daniel Unedo uploaded a video with no background music at all, foraging for salad ingredients in the wild. YouTube's copyright system sent him an automated claim anyway, triggered by birdsong in the background.

In 2015 the Australian artist Sebastian Tomczak uploaded a stretch of white noise and received five copyright claims.

Julia Reda has argued that cases like these show the overreach was happening already, in an era when tech companies were under no legal duty to police copyright. Once the EU's strict new law makes copyright protection an obligation, this can only happen more often.

By giving claims the backing of law, the new rules prepare the ground for copyright trolls. That is an invisible risk for everyone who makes derivative work.

Unedo's last video is five years old and his channel still has just over a thousand subscribers. The three-minute video that drew the claim has 110,000 views on YouTube.

Under it, Unedo wrote that the episode changed how he saw YouTube: he wanted to keep going but had lost the heart for it. "Google just ignored it and let copyright-troll companies bully small creators. It doesn't care about us. So I hope that in future I can publish my work on a platform that is fully open-source and belongs to everyone."

8. Filtering of speech spreads, beginning at the moment of upload

Filtering what people upload is something readers of Chinese websites know well.

Wikipedia founder Jimmy Wales, World Wide Web inventor Tim Berners-Lee, and others wrote an open letter arguing that "upload filters" would turn the internet from an open platform into a tool of "automated surveillance and control." David Kaye, the United Nations special rapporteur on freedom of expression, also warned about "censorship before publication."

Gus Rossi, global policy director at Public Knowledge, said in a statement: "Platforms large and small are likely to implement the directive worldwide, which will weaken users' ability to share memes and political satire online."

Companies and commentators are not the only ones worried. Jaroslaw Kaczynski, leader of Poland's ruling party, said much the same, that the EU's new copyright law "could threaten freedom of speech," and that Poland might not adopt it in full.

There is no shortage of examples of "protecting copyright" being used as a pretext to interfere with speech. Take Nando v Movies, a film-commentary channel on YouTube with 170,000 subscribers. His method is to rewrite a film's script into a better story — work that leans heavily on cutting footage from the film itself, and the films that need rewriting tend not to be good ones, like the DC superhero movies he loves to rework. Studios went to YouTube claiming infringement; the videos stayed up, but the ad revenue vanished. With no revenue there is no way to keep making them. Nando v Movies's videos now consist mostly of narration over a slideshow of stills, and are inevitably less compelling.

When speech is punished, people grow less willing to speak, and speak less.

You do not need a Soviet-era censor sitting in an office marking pages with a red pen. In the name of copyright, the right to speak simply goes away. Film studios can do it, and so can music companies, corporations, and individuals.

The copyright law has passed the European Parliament; each European country now decides whether to follow it. But a comfortable margin already tells you a good deal about where governments stand.

For the long-running conflict between internet platforms and content creators, this may not be the end of one problem but the beginning of many.

Cover image / banner image / visualhunt