Over the past year or so, the European Union's tangled web of triply-indirect democratic processes have been churning out a copyright directive. This has been vaguely reported on with semi-sensationalized names like the "link tax" and "censorship machines". In response to these names, creative industry lobbyists have alleged a vast net-wing conspiracy to mislead the public into believing that the upcoming EUCD will kill the Internet. I'm going to break down the two most controversial portions of this directive and explain what they do and what you should be angry about.
If you don't feel like reading the whole thing, here's the TL;DR:
- The "link tax" is a repeat of a failed idea that will result in more consolidation in certain markets by trying to put a price on free publicity. However, it might still be the best option for the news industry.
- The "censorship machines" don't explicitly call for automated enforcement, but leave service providers in the lurch as to what practices they need to do in order to avoid liability. It needs clarification.
- New common-sense copyright exceptions in the EUCD were either removed, refused for consideration, or are significantly inadequate and limited.
- France's government needs to stop giving me excuses to imagine them complaining about memes in a stereotypical French accent.
- The European Union legislative process is in significant need of reform.
The Link Tax - Article 11
EUCD Article 11 establishes a new exclusive right based around summarization and aggregation of news content. This is tailored specifically to the activities that Google's news search performs.
Is this a "link tax"?
Early versions of the EUCD were vague enough to be interpreted as establishing that sharing links would constitute copyright infringement. The language of the directive has been adjusted to be specific to news aggregators.
Wait, haven't we been here before?
Some time ago, Spain passed a copyright law which required Google license content from news outlets. From inconsistent reporting on the matter, I've heard that this law also established a price floor for aggregation rights, but I failed middle school Spanish so I can't read the law. Germany also passed a similar law.
The result was very predictable: Google took their toys and went home. Google News Spain was shut down and Spanish news providers were blacklisted from other Google News properties. In Germany, large news providers came to an agreement whereby they would grant free (or practically free) licenses to Google. The ultimate result is that outfits that didn't grant free license to Google (which were overwhelmingly smaller agencies) saw significant decreases in web traffic and total revenue due to not being aggregated.
Why is the European Union repeating the mistakes of their member states?
Presumably, to get a larger economic bloc to adopt the same strategy and export the law everywhere.
From an outsider's perspective (I'm American), it seems obviously doomed to fail. Google News is a public service, not a profit-bearing enterprise, and there isn't any revenue for Google to siphon off and pay news providers with. That's why they can so easily take their toys and go home. Things like GDPR work because it still makes sense to make some money rather than all of the money, and it's easier to comply everywhere than to only comply for EU citizens. However, when compliance is more expensive than leaving the market, the solution is always to leave the market.
However, we should take a moment to appreciate the situation the news industry is in. They are one of two industries that completely broke their business model on the back of the Internet. They assumed print circulation was enough, and that they could supplement that with advertising on freely posted articles. This didn't work out - Internet advertising has been devalued, and most people block it as a security measure now. Many news outfits are failing, and that's an existential threat to any kind of democracy.
From the perspective of the EU, a free and independent press is a necessity to a functioning democracy. That's free-as-in-freedom, as in the government can't restrict what they report on; and independent-as-in-money, as in their funding needs to come from disinterested parties. The first part is easy, but the latter requires a market value for news high enough to pay for it to be created. We can't have news funded out of the goodness of billionaires' checkbooks: the editorial independence of the Washington Post should not depend on Jeff Bezos' willingness to allow extortionists to publish his sexts.
What should the European Union should have done instead?
There's no good answer. The alternative would be to subsidize journalism. However, government funded media is a similar compromise of editorial independence. Do you want the American system, where the media believes whatever the President says (at least, when he's not blatantly wrong) in exchange for access to slightly easier reporting? Trying to price-fix the market is the least damaging option.
Of course, doing nothing is also an option. The current industry solution is for news outfits to implement paywalls. However, the free traffic and shavings of a penny in ad revenue obtained from news aggregators is still valuable enough that a hard paywall is not an option. Instead, many outfits have adopted graduated paywalls where the first handful of article loads are free and then cost money after, say, the first five in any given month. This appears to have improved the fortunes of many journalistic outfits, but it doesn't work for everyone. There's plenty of smaller outfits that benefit more from free traffic from search engines than they get from charging money.
In other industries that didn't see the Internet coming, like music, the solution was to implement a subscription service with everything imaginable on it. This is not without it's problems, however. First off, the amount of money you pay does not vary based on how much you listen. This means that artists are paid by taking the total revenue of the service (after Spotify's cut) and dividing it by how many plays they got. This, obviously, completely disadvantages niche artists. Furthermore, Spotify is not completely divorced from the mechanics of advertising, because they have a free tier. This further adjusts their incentives in the wrong direction. From Spotify's perspective, their goal is to get and retain subscribers, which means they need to give free tier users what they want, but just annoyingly enough to get them to subscribe. Of course, many music listeners have fairly limited tastes in music, and it's not in Spotify's best interest to broaden people's tastes in music. If this sounds like the ClearChannel era of radio, then you're right: Nothing has changed and music is still terrible.
The news industry is sort of in a similar position, except their subscriptions are per-outfit. The only thing that would have to change is some kind of "Spotify for News" that gets you all the news subscriptions at once and divvys out payments accordingly. I don't know if this would be any better for news, except that it may at least get a few more people to start paying for it.
The Censorship Machines - Article 13
EUCD Article 13 eliminates the mere-conduit "safe harbor" status that service providers enjoy to disclaim copyright liability. Instead, service providers are required to either license their users' infringing content or provide a reasonable and effective method for preventing infringement on their platform.
Wait, that seems reasonable.
Early drafts of Article 13 contained a recital explicitly stating that the purpose of it was to force social media platforms to implement automated copyright enforcement mechanisms. Furthermore, the scope of Article 13 has been reduced, with services like Wikipedia and Github being excluded from enforcement requirements. The anger over Article 13 started with these earlier, stricter drafts.
Are these "censorship machines"?
YouTube already implements an automated copyright enforcement mechanism called Content ID. After many years of policy tweaks, it almost works as a fair way to keep blatant infringement off of YouTube while still allowing fair use. If Article 13 in practice mainly served to require video services to adopt similar systems, it probably would be tolerable. Not acceptable, but tolerable - Internet society as we know it today could still exist at the same scale.
There are a few problems with this interpretation, however:
- YouTube management does not believe that YouTube Content ID currently complies with Article 13.
- Licensing every piece of content ever is impractical. While it's mentioned in the law, licensing is not always available, so we need to consider what happens when you can't get a license.
- The requirements for filtering are based on "high industry standards of professional diligence", and how that's interpreted can vary wildly. Since the EU will become a fully civil law regime after Brexit, it's odd that they're relying on courts to decide what that means. Jurisprudence doesn't go as far over there as precedent does over here.
The problem of filtering is not a problem of technical ability. Software exists to find similarities between images, videos, and audio patterns; this the current state of the art with regards to automated copyright enforcement. However, content similarity is not the definition of copyright law. It won't catch most unlicensed derivative works - though I don't think that's entirely the intent of this directive to begin with. There is also no provision to establish a filtering database. As it currently stands, things like YouTube Content ID's rights database is proprietary Google information. So I can only assume that the responsibility to filter starts when a copyright owner hands a social media service content samples and says "Make it happen", which seems like an extra burden for them. Remember: The reason why this whole thing is changing is because manually filing takedowns is a total pain in the ass.
Furthermore, copyright contains many exceptions that social media platforms are now expected to adjudicate, despite not being legally empowered to do so. (In America, these are called "fair use", but it's not called that in other countries.) In other words, if you think something is covered under exceptions to copyright and allow it, and a court disagrees, you're in the hook for the infringement. I didn't see anything in the directive or the trilogues covering what happens, aside from a general provision that non-infringing content should not be falsely removed from social media platforms. Given that the cost of being too loose with exceptions will likely be far higher than the cost of mistaken takedowns, the burden is almost certainly going to be on social media users to justify the copyright status of their work, rather than on claimnants to justify their ownership of other people's work.
Okay, so what does YouTube have to do to stay operating in Europe?
It's unclear, at least until this directive is implemented by member states. Looking through the trilogue compromise text, the opinion of the people writing the directive is that it would be the opinion of the government (probably through the courts) whether or not a particular set of measures implemented by a social media platform were adequate at stopping infringement. I personally believe something like YouTube Content ID is good enough for Article 13, but YouTube themselves doesn't.
Here's an example of how a court case might proceed in an EU member state after Article 13. Facebook launched a video platform a few years ago which was notorious for infringing the copyright of YouTube users. If they had launched in a post-Article 13 world, YouTube would be able to demand Facebook implement an equivalent to their own Content ID system, and if they refused, Google would be able to sue for so much money that the European Commission would have to propose and fast-track a new competition law to keep Google from buying Facebook as a settlement condition. YouTube would likely be held as the standard all video platforms must adopt to avoid copyright liability.
It's important to note that the directive appears to be written to allow unspecified actors to increase filtering standards whenever appropriate. If a new kind of machine-learning software is created that can accurately predict the results of court cases and adjudicate law, it's likely that the EU could mandate it's use.
Is this a "meme ban"?
It is the position of the European Commission (or at least their Twitter account) that memes are protected under exceptions to copyright. (However, please see the below section on "The Missing Exceptions".)
Concerns that memes would be accidentally banned by automated enforcement are the same issue as the "censorship machines".
Google was able to get a license from GEMA, why can't they do the same elsewhere?
The problem with telling social media companies that they need to license content isn't that they can't negotiate with collective licensing agencies. If the law was preferential towards collective licensing, that probably would be more workable. The problem is that it isn't; every work imaginable must be licensed or filtered. There is no legal provision specifying how many licenses Google needs to buy, just that they need to buy them, or implement filtering that an EU court agrees is "good enough".
Yes, the law was written with collective licensing societies in mind, but not all works are available through them. It is not legally possible to license the entire body of copyrighted works. If I shit on a piece of toilet paper, that constitutes a new copyrighted work that I can demand YouTube license or filter.
The Missing Exceptions
Previous versions of the directive included additional provisions for new exceptions to, or limitations upon, copyright:
- Research institutions are allowed to mine copyrighted text
- Distance learning programs are allowed to use copyrighted material for educational purposes
- Cultural heritage organizations (museums) are allowed to republish out-of-commerce works
- Freedom of panorama (the right to take pictures of copyrighted buildings) is now EU-wide
- Visual reproductions of public domain works are not recopyrighted
Wait, why did you say missing?
So, the freedom of panorama is probably gone. It's not even mentioned in the trilogue text. I've heard rumors that some MEPs plan to push for amendments to bring it back. I don't believe that would be successful. Doesn't that just have to go through the EC Trilogues again? I mean, a lot of France's creative industry was absolutely pissed that the Culture Committee even recommended adding a specific exception for user generated-content.
This is where I have to start complaining about the European Union's really, really dumb way of writing laws. It's not just that the MEPs vote on a thing and it becomes law, like in almost all of the member states of the EU. Instead, it goes to a thing called the Council of the European Union, which then brings the European Commission along to go yell at the MEPs who voted for it until they agree on something else. These things called "trilogues" are often used as an antidemocratic process by which unpopular laws and provisions to law can be forced into existence. Publicly oppose the thing you want, push for it in private in the Commission, then go to the voters and complain that it's the EU's fault. It's unfair for everyone involved.
The EU needs significant reform and cannot act as an international cudgel for Spain and France's current governments to bully the rest of the union into giving them what they want. For those who aren't aware, the EU has a bicameral legislature, but the elements that are most directly responsive to democratic concerns are deliberately withheld from power. Only the European Commission can introduce new directives (usually), and nothing passed by the European Parliament can come into force without the approval of the Council for the European Union, whose membership is determined by political appointees from member states. The Council for the European Union (not to be confused with the European Council) operates by a complicated "qualified majority" system. Directives only pass the CEU if the following applies:
- Approval of member states who represent at least 65% of EU voters
- Approval of at least 55% of member states, except when the EC didn't introduce the directive, when this arbitrarily increases to 72%
Effectively, the EU runs on a three tier system. The European Commission, the least democratic, gets to determine what is voted upon. The Council for the European Union, which is slightly less undemocratic, can override the EC, but arbitrarily requires a supermajority. And the European Parliament - the thing that EU citizens can actually vote for - only has veto rights. This is reminding me of the time back when Americans couldn't vote for Senators and the US Senate effectively served as a way to shut down the will of the people. The Commission and the Council should go away and the Parliament should have sole legislative power within the EU.
Gee, Brexit can't happen fast enough!
The UK has passed their own share of terrible copyright laws. This is the country that gave us the "Home Taping Is Killing Music" nonsense from decades past, after all. Back in 2016 they almost managed to get an automated enforcement provision for search engines, and in 2010 they mandated three-strikes on ISPs.
The UK is also not much better in the democracy department. First off, it is still technically an absolute monarchy and a theocracy; however, this is mostly theoretical as the crown has ceded most of it's power to parliament and mainly serves as a source of revenue from royal lands. However, the UK Parliament is also bicameral, and it also has a similarly antidemocratic second chamber called the House of Lords. Except instead of representing countries within the United Kingdom, they don't really represent anything. Much of the House of Lords is appointed by the Prime Minister, or are religious leaders (because theocracy), or are there literally by birthright.
The underlying problem driving the EUCD is that acts, laws, bills, and directives that are good for the state are not necessarily good for the people. This problem is not unique to Europe and every country has some kind of antidemocratic procedures somewhere. The solution is to identify and reform those procedures. Furthermore, I refuse to give credit to Brexit supporters for potentially sparing the UK from the EUCD. That is an accident; their intended goal was to close the border to EU citizens.
But doesn't being outside the EU mean you won't have to deal with this law?
First off, it's difficult to only comply with the law for the EU. Social media platforms are going to be under enormous pressure to implement EU-mandated copyright filters worldwide.
Second, the Council of the European Union is not the only undemocratic institution by which countries can demand new copyright law. Once implemented in the EU, it is likely that there will be a renewed push for similar laws in other countries, such as the US, or Japan. Once enough countries start to implement such bills, it will become a point of contention in international trade negotiations. If you thought the trilogues were dumb, wait until you find out about how the WTO and WIPO works.
Oh, by the way: Exceptions to copyright don't work the same way. In fact, the "cultural heritage organizations" exception I listed above has a massive carve-out. You only get to use it if the work originated in the EU, because anything else would constitute a huge violation of many, many international trade treaties. It's far easier to export more copyright law than less.
Why the hell does the EU always prostrate itself for the copyright industry?
Because it has them. Copyright isn't an American invention. Any country with a large creative industry will find it's legislative priorities dominated by the need for stricter and stricter enforcement. Japan, for example, has a huge creative industry and an absolutely extreme copyright code to match. The EU is no different; it has several member states with large and active creative industries.
Do you want to know what the EU doesn't have? Technology companies - at least, not nearly to the degree that America has. America has to balance it's copyright laws against the interest of the technology industries that we completely dominate. If the EU wants to compete with America, it makes more sense to protect the industries it has - that is, creative industries - rather than to try and grow a technology sector in a world where Apple, Google, Facebook, Microsoft, and Amazon are the technology industry.
Remember what I said before about what's good for the state isn't always good for the people?