On September 12th this year, the European Parliament voted on and approved a sweeping copyright directive that aims to create a unified framework for handling copyrights and licenses within the EU. The proposal is currently passing through the Trilogue, expected to conclude in January 2019. If formalized, EU member states would be required to pass laws to enact this directive. And if that happens, it might drastically change the media landscape and possibly even how the internet works for Europeans.

Specifically, there are two articles in the directive that have drawn massive criticism: Article 11 and Article 13.

Article 11: The “Link Tax”

On the face of it, the intent of Article 11 is fairly benign and laudable: give publishers the right to ask for paid licenses when their news stories are shared on commercial aggregation platforms (like Google News).

Right now, publishers first have to get the copyright assigned by the authors, and then have to prove ownership before being able to assert any copyright claims. Under this article, publishers automatically have the exclusive rights for what they publish for one year, making it easier to assert their claims. Platforms would then have to have licensing agreements with publishers before being allowed to excerpt and link to articles.

The trouble starts with the vague wording. The article does say that a mere hyperlink or “individual words” aren’t subject to this, but when do “individual words” stop being individual and become an excerpt that has to be licensed? Furthermore, the article doesn’t clearly define what constitutes a commercial platform¹. Would RSS readers count as “commercial platforms”? What about a popular Facebook page?²

There’s actually even a precedent for this article: Both Spain and Germany have already enacted very similar laws in their countries. And in both cases, the effect was the same: Google simply closed down Google News³, causing web traffic to drop for publishers, especially hurting small ones and even forcing some to close as a result. In Germany, publishers eventually bent the knee and begged Google to resume its aggregation. The reason was the same in both cases: Google and Facebook can and do control who gets access to their audiences, and so the power lies with them, not the publishers. Publishers, on the other hand, who are used to being more or less the only game in town, now face the problem that basically anyone with access to a computer and the internet can basically do the same thing as they do. They are no longer in control of their readership.

Nothing exemplifies this more than that there’s already a technical solution for this exact kind of thing publishers want: The robots.txt file. If a news organization doesn’t want Google (or any other aggregator) to link to their stuff, they could add it to the robots.txt file⁴. If News organizations really wanted to stop Google from linking to their articles, they could have done so at any time in the past. It’s telling that none ever did, and instead lobbied for this article instead to force Google et al. to give them money. More likely, they will just continue shooting themselves into their feet, especially if platforms like Facebook decide that simply linking to publisher’s articles isn’t worth the liability, and just disallow such links altogether.

Article 13: The “Meme Ban”

Article 13, like Article 11, sounds very sensible: Platforms are liable for copyright violations by content uploaded by their users.

Currently, platforms are basically exempt from having to license user-uploaded content until they are notified of a copyright violation⁵, upon which they have to respond in a timely manner. This system is biased to false negatives: Content can be uploaded that shouldn’t have been, and action is required from the copyright holders to remove it. Copyright holders cannot sue the platforms for hosting such content if the platforms follow this process, and can only sue the uploader.

Under Article 13, platforms would have to proactively check for copyright violations, or ensure that they have all the required licenses beforehand. If they don’t want to or can’t do that⁶, platforms face the risk of being sued by the copyright holders. This system is biased towards false positives: platforms block any content that might be copyrighted, due to fears of liability (and associated fines). Copyright holders don’t have to do anything (great for them!), and all the burden is put on the platforms to ensure that they do not host unlicensed copyrighted works.

Again, on the face of it, this sounds sensible. If someone uploads a movie to YouTube, then why does the movie’s studio need to be on the lookout for such uploads in order to tell YouTube to remove them?

The first problem is that copyright law differ from country to country, so deciding what’s legal or not is difficult. The second problem is that platforms have to use automated systems to scan uploaded content for copyright violations due to the massive amounts of data they get (In 2015, about 500 hours of video were uploaded to YouTube every minute). And such automated systems, no matter how sophisticated they are, will be incapable of distinguishing between fair use (for example, a video critique of a blockbuster movie that shows excerpts to illustrate its points), and copyright infringement. With memes, it’s even more complex: Many countries do not exempt them from copyright laws in the first place, and even if they did, memes are highly dependent on the context they’re used in. No AI today can reliably correctly categorize such uploads (a general AI might be able to, but such an AI could result in other, much more severe, problems for us (YouTube)).

Like with Article 11, we already have examples of what would likely happen if such systems became mandatory: when musician James Rhodes uploaded a video of him playing music written by Johann Sebastian Bach on his own instruments, Facebook muted the audio in his video, since Sony Music claimed it owned the rights to Bach’s music⁷. Sony Music even stuck to its guns through several appeals, only relinquishing its claim after mounting public pressure. Under Article 13, Rhodes likely would not have been able to upload the video in the first place, and unlikely be able to appeal that decision (platforms have no obligation to host your content, and would optimize to avoid liability instead).

Or YouTube’s Content ID system deciding that birds chirping in the background are a copyright violation (no, really). Or that motorcycle engine sounds are also copyrighted. The upshot of the current system is that creators are able to fight those claims. Under the proposed article, platforms would likely just deny the upload in the first place, giving creators no recourse to appeal that decision, since the platform would face the liability if it turns out that it was copyrighted content.

And finally, such systems would then have to rely on copyright holders to correctly claim their works, opening the doors for trolls who could claim copyright for things they don’t actually own. Platforms are more likely to do whatever avoids liability instead of fighting for their user’s fair use rights.

So, What Now?

In their current form, both Articles have good intentions, but are so confusingly or vaguely worded that their actual effect can be hard to predict. They are clearly aimed at the internet giants Google (including YouTube) and Facebook, but are worded in such a way as if the internet only consisted of these companies (GDRP has a similar problem).

And while proponents argue these articles ensure fair compensation for content creators, critics fear they give too much power to copyright holders and copyright trolls, and will prevent new platforms from starting up due to the associated difficulty in implementing such automated systems.

There are also other articles in this directive whose effect is still very much unclear. For example, Article 12 gives sports leagues exclusive rights to any media of a sporting event, possibly including fan-created photos and videos.

In the end, the existing big platforms won’t be really harmed by it. They have the resources to build and maintain any automated systems required to avoid liability anyway. It’s going to be the small creators who depend on the platforms to host their content who will lose out, since they will face a very steep uphill battle to get anything uploaded if it even contains a hint of copyrighted material (and sometimes even if it doesn’t). In fact, big platforms might even profit from the directive in the end, since any rival upstarts will be dead on arrival due to the cost of implementing such systems. Copyright laws need to be updated for the age of the internet, but this isn’t it.

No Platform, No Money

De-platforming refers to banning someone from a platform that they rely on to spread their message. It has mostly happened to right-wing figures, like Alex Jones, or Milo Yiannopoulos. And whenever they are de-platformed, they claim they’re only getting stronger, and that it won’t work. But Milo Yiannopoulos’s collapse shows that no-platforming can work.

Music Magic

One of the more magical apps on smartphones are music recognition apps like Shazam. While we humans are great at recognizing songs we already know, they can be a godsend if you’re in a noisy place, and you just want to know the name of the song that’s currently playing: How Shazam Works (YouTube)

Sounds of Mars

You’ve probably seen quite a few photos of Mars by now. But the recently landed Mars Insight has something that no other probe had before: instruments so sensitive they can double as microphones: Hear the First Sounds Ever Recorded on Mars.

📖 Weekly Longread 📚

As the games industry gets more and more crowded, Polygon looks at what small teams are doing to stand out: There are too many video games. What now?

🦄 Unicorn Chaser 🦄

“This is possibly the most British thing I’ve ever seen an animal do”: Watch How Hermit Crabs All Line Up to Exchange Their Shells