Not Even Hiding It Any More: EU Council Explicitly Pushing For Mandatory Upload Filters

One of the key talking points for supporters of Article 13 in the EU Copyright Directive is to absolutely deny that it requires mandatory upload filters. Of course, as soon as you ask them how an internet platform could possibly abide by the rules of Article 13 without implementing mandatory upload filters, they suddenly change the conversation. Usually to something about how YouTube is ripping off all musicians. This is... weird. First of all, YouTube already has its giant upload filter in the form of ContentID. Second, if they can't tell you how it doesn't require upload filters, then... it requires upload filters.

As the trilogue negotiations continue between the EU Council, the EU Commission and the EU Parliament, the Council has apparently decided to drop the pretense and is now explicitly demanding mandatory upload filters. The newly proposed language says that any site is liable for all infringement committed by their users unless they block any infringing works they've been informed about from ever appearing on their sites again. It's a "notice and stay down" requirement -- which has all sorts of problems. First of all, this assumes that every use of the same work is equally infringing. It does not take into account that one use may be infringing, while another may be fair use or fair dealing. Second, it requires incredibly expensive technology. ContentID already cost Google over $100 million... and it's not very good. Tons of stuff still gets through. So now, basically, any successful smaller platform would have to spend ridiculous sums of money to implement a useless filter that won't work... and when things slip through, they're still liable for massive damages.

And, notice what's missing? What happens if these filters take down content they should not? This happens all the time. But here, of course, there is no punishment for false notifications or for mistakes. While the Council tries to get around this by saying the rules "shall not affect legitimate uses, such as uses under exceptions and limitations," that's entirely meaningless. How the hell do you train a filter to understand parody? Or fair use? Or any other limitation or exception? Google has spent $100 million on its system and it has no clue how to determine fair use.

The link above to Julia Reda's site has more info on the current state of the negotiations, but suffice it to say that this still appears to be an utter disaster for the internet, as you have people who have no understanding at all how the internet works, passing sweeping regulations that will have massive consequences for speech online.