Thank you for your comments on Twitter welcoming my feedback on the EU’s proposed copyright reform. I’d like to discuss in particular Article 13, “Use of protected content by information society service providers storing and giving access to large amounts of works and other subject-matter uploaded by their users”.
Firstly I should say that I am in general in favour of the EU’s overall approach to data collection and privacy, and of the goal of copyright reform. I am also in favour of the goals of Article 13: improving the rights of creators over giant American tech companies (and there was a great recent letter in the Guardian on this point). My concerns with Article 13 are not with what it’s trying to do, but in the implementation: it is loosely worded to the point where it will cause serious problems, it materially underweights the costs and hassle of compliance, and it pushes for unreliable technology without requiring any oversight.
I should unpack those assertions a bit.
Content ID, the approach of using technology to detect copyrighted material, doesn’t work very well. Or, more accurately, it’s way too easy for rightsholders to incorrectly assert rights and it’s way too hard to appeal because an algorithm doesn’t listen to explanations. I’ve experienced this myself: in 2012 myself and my dad drove around our home town and videoed all the streets to provide a historical artifact that people could look at thirty years from now to know what the town was like. I found a piece of “elevator music”, available under a Creative Commons licence, attached that to the video, and uploaded it to YouTube. And then later I received a copyright claim from a record company stating that I’d used music from their catalogue. To be clear here: I had not. They had misidentified the music as being another song, one over which they hold rights. But to explain this I have to file a “counterclaim”, hand over my home address, think about naming a solicitor… or if I don’t do that, Google put a “copyright strike” on my account and start removing my ability to earn money from a thing I made. In this… I’m a creator. And I’m being forced away from my creation. It’s a small example, but it’s characteristic: laws designed to protect creators (which, again, I’m in favour of) need to ensure that they protect the rights of all creators, not just the ones who have a legal department and millions in the bank to fund an indiscriminate barrage of automated takedown requests, especially when those takedown requests are wrong.
But even if we assume that the technology is 100% reliable (ha! have you ever seen technology that was?), the wording of Article 13 is still problematic. Essentially it says “if you allow people to upload things, you need to implement Content ID now”. The text refers to those who store and provide “large” amounts of works, but in no way define what that actually is! I run a discussion forum for a podcast, and there people can post messages and, importantly, also post files. Are we hosting a “large” amount of works? I doubt it, but… who knows? And how do I implement Content ID anyway? In order to tell whether an uploaded thing is copyrighted I’d need access to every copyrighted work ever; this inevitably means that I’d need to subscribe to some sort of “is this thing copyrighted?” third-party service, which will not at all be free of charge and would be a burden to integrate even if it were. This is a major pain to do, and it essentially applies to everyone because it’s not at all clear where the dividing line is for “you are now big enough to have to do this”.
I get the goal of this. I sympathise with the goal of this. It’s to say: hey, huge hosting sites, YouTube et al, you have to try harder to stop people uploading music and movies without ensuring that the creators get paid. You can’t hide behind the excuse that it’s up to the uploader to check and you’re just a neutral repository; you bear some responsibility here, so get on with it. I think that is correct and proper! And I also sympathise with the idea that defining “huge hosting sites, YouTube et al” in actual legal terms is pretty difficult. But the current wording is a poor attempt at it.
It’s like… well, imagine that we were worried about stabbing incidents in nightclubs, and so we want to pass a law saying that everyone going into a club needs to be searched to ensure that they don’t have a knife. But the law says “all persons entering a place of assembly with music must be searched”, and then someone who runs a coffee shop on the High Street says “hey, surely I don’t have to search everyone who comes in to buy a latte, do I?”. And they do, because the description probably applies to them too… and now they’re turning customers away because they don’t want to be searched in order to buy a cup of tea. The goal of the law is a good idea; the intentions behind it are worthwhile. But it’s drafted in a way that pulls in a whole load of people that it’s not aimed at, and fills those people with uncertainty and concern. If you’re a multinational corporation then licensing some sort of Content ID service is just one more cost on the balance sheet. If you’re not, it’s a huge deal. Just the uncertainty of whether you might be on the list is a huge deal.
And saying “you probably don’t count as ‘large’, so don’t worry about it; we promise to only prosecute those people who deserve it” is not at all encouraging. Especially if you’re a small host and the people who’ve decided to sue are a huge media company whose software has misidentified your theme tune as one of their records. There’s a huge incentive for content owners to threaten small hosts under this law, and it’s really hard to fight back against that; if I get a letter saying “you are a host distributing content without having a content scanning policy as per Article 13” then I can’t afford to even argue about that; anything involving the courts is in itself expensive and time-consuming. I’m not going to instruct a solicitor to help with that case, at a hundred pounds an hour; I’m going to shut my forum down. And this applies even if that letter I received was sent by an algorithm which misidentified some music and no human ever saw it. If I can say “that law doesn’t apply to me”, then I’m good. If I can only say “I don’t think it applies”… then I’m sunk.
Smaller hosts and creators are not, and really mustn’t become, necessary fallout in a fight between big content creation firms and big internet companies. Please don’t let this law go forward in the way it’s currently written; have it do the job it’s intended to and not more.
Many thanks for taking the time to read all this. I look forward to your response, and to have further discussion about it all. I’ve posted this letter to my website as well so that it might help others also looking to contact their representatives; I’d be happy to update that with your thoughts as well (or not, of course, if you prefer not).