Top EU Court's Adviser Regrettably Fails To Recommend Throwing Out Upload Filters, But Does Say They Should Block Only "Identical" Or "Equivalent" Copies

from the out-of-their-tiny-minds dept

One of the last hopes of getting the EU’s terrible upload filters thrown out was an intriguing legal challenge brought by Poland at the region’s highest court, the Court of Justice of the European Union (CJEU). As is usual in these cases, a preliminary opinion is offered by one of the CJEU’s special advisers. It’s not binding on the main court, but can offer interesting hints of what the final judgment might be. Unfortunately, in his analysis Advocate General Saugmandsgaard Øe recommends that the CJEU should dismiss the action brought by Poland (pdf), because in his view Article 17 of the EU Copyright Directive is compatible with freedom of expression and information.

That’s a huge disappointment, since many hoped he would unequivocally rule that upload filters breach fundamental rights. However, the Advocate General’s opinion is by no means a complete disaster for users of online sharing services. He recognizes the right of people to make “legitimate use of protected subject matter.” Specifically, that means people must be able to rely on the EU’s exceptions and limitations to copyright. Moreover:

In order for that right to be effective, providers of such [online sharing] services are not allowed to preventively block all content reproducing the protected subject matter identified by the rightholders, including lawful content. It would not be sufficient for users to have the possibility, under a complaints and redress mechanism, to have their legitimate content re-uploaded after such preventive blocking.

This is a huge point. It means that copyright companies cannot demand that upload filters block every use of their material, since that would prevent legal transformative uses such as memes, parodies, commentary etc. Saugmandsgaard Øe concludes with the following observation to the CJEU:

Consequently, sharing service providers must only detect and block content that is ‘identical’ or ‘equivalent’ to the protected subject matter identified by the rightholders, that is to say content the unlawfulness of which may be regarded as manifest in the light of the information provided by the rightholders. By contrast, in all ambiguous situations — short extracts from works included in longer content, ‘transformative’ works, etc. — in which, in particular, the application of exceptions and limitations to copyright is reasonably foreseeable, the content concerned should not be the subject of a preventive blocking measure. The risk of ‘over-blocking’ is thus minimised. Rightholders will have to request the removal or blocking of the content in question by means of substantiated notifications, or even refer the matter to a court for a ruling on the lawfulness of the content and, in the event that it is unlawful, order its removal and blocking.

Crucially, this says that unless it is absolutely clear-cut that there is copyright infringement — because an identical, or equivalent copy is uploaded — user uploads must not be blocked by default. Instead, a more detailed complaint must be made by copyright holders, possibly involving a request for courts to rule on the legality of a transformative use. That’s very far from what those pushing for upload filters want, and represents a major limitation on the latter.

It’s an obvious compromise position, and as such could well be adopted by the CJEU when it hands down its definitive judgment at a later date. Saugmandsgaard Øe says that yes, upload filters are acceptable in the EU, but can only be used to block identical, or near-identical copies. In his full opinion, he also affirms strongly and repeatedly that other legal uses of copyright material must not be blocked by upload filters. And there’s a nice sting in the tail of his analysis. In a Postscript, the Advocate General comments on the European Commission’s recent “guidance” to national governments on how they should implement Article 17. As Techdirt noted last month, this guidance introduced a huge loophole that would let copyright companies “earmark” any upload that they claim “could cause significant economic harm”, even if likely to be a legitimate use of protected subject matter. Earmarked uploads would lack key legal protections, and Saugmandsgaard Øe is having none of it:

If this is to be understood as meaning that those same providers should block content ex ante [in advance] simply on the basis of an assertion of a risk of significant economic harm by rightholders — since the guidance does not contain any other criterion objectively limiting the ‘earmarking’ mechanism to specific cases — even if that content is not manifestly infringing, I cannot agree with this, unless I alter all the considerations set out in this Opinion.

This is basically Advocate General-speak for “you must be out of your tiny minds”.

Follow me @glynmoody on Twitter, Diaspora, or Mastodon.

Filed Under: , , , , ,

Rate this comment as insightful
Rate this comment as funny
You have rated this comment as insightful
You have rated this comment as funny
Flag this comment as abusive/trolling/spam
You have flagged this comment
The first word has already been claimed
The last word has already been claimed
Insightful Lightbulb icon Funny Laughing icon Abusive/trolling/spam Flag icon Insightful badge Lightbulb icon Funny badge Laughing icon Comments icon

Comments on “Top EU Court's Adviser Regrettably Fails To Recommend Throwing Out Upload Filters, But Does Say They Should Block Only "Identical" Or "Equivalent" Copies”

Subscribe: RSS Leave a comment
21 Comments
This comment has been deemed insightful by the community.
Jojo (profile) says:

Disappointing, but still promising

Even if the recommendation is disappointing, I’d do think that this is still a technical victory for us. Why you may ask? Well the recommendation creates a possible situation which forces Safeguards as a requirement instead of an option when considering Upload filters. This is a potential blow to Pro-Copyreich state members since this means that their laws may have the possibility of being thrown back to square one. It does make the law-making process ten times more messier.

It’s true that I would’ve liked it if the AG have thrown out Art.17 entirely, but the situation could’ve been so much worse. I just hope that Poland’s lawsuit survives though.

Anonymous Coward says:

Creating an upload filter that proves two files are identical is easy.

Creating one that proves two files are equivalent is extremely hard, and, with today’s knowledge, impossible to implement with 100% accuracy.

I really wish I could understand how these politicians expect these things to be done and work in the way they describe. The only rationale I can come up with is that they’re not talking to or not listening to the experts who would end creating these filters.

TaboToka (profile) says:

Re: Re:

I really wish I could understand how these politicians expect these things to be done and work in the way they describe.

Think of it as the spawn of the unholy union of The Bike Shed Effect and Dunning-Jkruger Effect:

  1. I am familiar with computers and programs (the kids call them “apps”) because I use them all the time
  2. I know they can do many spectacular things, even think, as depicted in TV/Movies
  3. I know that people smarter/more skilled than me can write the apps
  4. Therefore, it must be possible—and easy—for those in #3 to do #2 and to stop people from uploading copyrighted stuff, but not memes using that copyrighted stuff , or those soundtrack mashups on the TikToks my granddaughter sends me
sumgai (profile) says:

Ya know….

From an American standpoint, this really is the best compromise possible. I mean, if Poland loses all the way across the board and Art. 17 stands, most of the world will see even harsher restrictions on uploading, or to put it another way, restrictions on communicating whatever idea strikes your fancy at the moment. But without any filtering at all, the WWW will see lawsuits like never imagined before, from all of the copyright maximalists. ("Your filter isn’t strong enough, it’s not keeping everything down. We’re suing you for facilitating infringement!")

So the hope would be that America can be dragged into a very similar scenario, where a C.M. must go to court EVERY TIME they want to take down a meme, or a satirical piece, or a critizism, etc. That puts the financial burden where it belongs, right back on the party desiring redress. Not to mention, lawyers like Biss et al will go broke, because thin-skinned dickheads won’t be able to hire them to harass cow-themed Twitter feeds. That’s a WIN right there!

Anonymous Coward says:

i don,t see how it,ll be possible for small startups or websites to filter all content,
audio, video, images, memes etc
and tell is this infringement , is it parody, fair use , legal use of ip.
eg film reviews which show short video clips .
it might happen that only google and facebook will have the tech and the staff to install such complex filters.
Many eu websites will simply chose to block all user uploads .
of course most politicans dont understand how tech works they are simply passing laws that are fovourable to big corporations and legacy media companys

TaboToka (profile) says:

The whole thing is stupid and ignores the obvious

The easiest thing to do is not allow anyone in the EU to upload anything.

Places like Techdirt don’t have the resources to implement even passable filers, so they will just geo fence the EU. Maybe places like TikTok and Insta will have to figure it out (as it is their business model) but they have enough € to hire lawyers and devs.

Usually, the easiest/cheapest solution that doesn’t suck too bad wins.

This comment has been flagged by the community. Click here to show it.

Dennis87 says:

dunkinrunsonyou

from the who-said-it deptThis week, our first place winner on the insightful side is Bloof with a response to someone making the plainly silly claim that Section 230 was created specifically to induce acts of censorship of conservatives by social media companies:Section 230 became law in 1996. Facebook was founded 2004, Twitter in 2006, YouTube in 2005, Google in 1998, MySpace in 2003

https://www.dunkinrunsonyou.net/

tp (profile) says:

Caveat: These are all identical copies...

1) if the file extension is the same, i.e. all .mp4 files are the same content
2) if the file name has any commonality, i.e. madonna.mp4 is the same as madonna.mov
3) if both content items have green pixels anywhere in the screen
4) if both content items have red pixels anywhere in the screen
5) if both content items have blue pixels anywhere in the screen
6) if both files file size absolute difference is less than 10k
7) if the urls are pointing to the same domain, i.e. http://foobar.com/madonna.mp4 is the same as https://foobar.com/madonna.mp4

These tests for "identical or equivalent" content are trivial to implement in a computer and testing it is quick enough to work for millions of content items.

PaulT (profile) says:

Re: Caveat: These are all identical copies...

So…

  1. You’re saying that content isn’t the same if you encode it in a different format?
  2. You’re saying that content isn’t the same if you rename it, and it’s not possible for different content to share a title?
    3-5. Erm, every colour has those pixels. It’s literally how pixels work.
  3. You’re saying that different content can’t be the same size?
  4. No shit, the content is the same if you’re literally pointing to the same file.

Yet again, the world’s worst developer demonstrates that he know absolutely nothing about his very profession.

PaulT (profile) says:

Re: Re: Re:2 Caveat: These are all identical copies...

I’m assuming you got autocorrected there and you meant markup and not mario up?

Basically, markup is a term from print formatting that’s used to refer to online display formatting (the M in HTML stands for markup). Whereas, markdown is a specific instance of a markup language, as far as I know the name is simply a pun.

Lostinlodos (profile) says:

Re: Re: Re:3 Caveat: These are all identical copies...

Yes, auto fuck up…
And what the hell apple: Mario doesn’t up, he 1Ups!

Koolness.
I figured it wasn’t different. But every now and then someone says “down”.
Ps, pdf, html, and bbcode are all up.
Never got the “down” aspect.

Learn something new, every day.
Thanks PaulT.
(Despite my annoyance with some of your political positions, I find you well versed in commons).

Scary Devil Monastery (profile) says:

Re: Caveat: These are all identical copies...

"These tests for "identical or equivalent" content are trivial to implement in a computer and testing it is quick enough to work for millions of content items."

As if we needed further evidence that you’re a technological illiterate.

Short answer; No, that doesn’t work. Thousands of false positives a day beg to differ with your armchair argument.
Meanwhile, as youtube uploaders have shown quite a lot, it’s childishly easy for anyone who deliberately wants to circumvent that filter to push an upload through. Cut out ten 1/10 second instances from an uploaded movie, selectively reduce frame rate or resolution. Displace the sound track with 1/10th of a second…

Once you’ve managed to train contentID to recognize every possible configurationm of a file the definition is wide enough that everything else will be caught in it.

The definitions you gave are all worthless and easily changed.

tp (profile) says:

Re: Re: Caveat: These are all identical copies...

it’s childishly easy for anyone who deliberately wants to circumvent that filter to push an upload through

This is why the copyright laws/DMCA wants companies to implement their repeat-infringer policies properly. I.e. once uploaded, the infringer has no defense for the repeat-infringer policy violations, and the account will be closed once the youtube admins manage to reach those videos. Getting the upload through is only the beginning, and keeping your accounts up and running while posting videos when companies are implementing their repeat-infringer policies. Manual checking is just ok practice for checking copyrights. if contentID doesn’t catch everything, a building full of video reviewers can easily catch those pirates.

Scary Devil Monastery (profile) says:

Re: Re: Re: Caveat: These are all identical copies...

"This is why the copyright laws/DMCA wants companies to implement their repeat-infringer policies properly."

You mean that once ContentID falsely recognizes a video or a DMCA complaint causes an unwarranted takedown thos uploaders should be banned for good? Because that’s the result you envision here.

"a building full of video reviewers can easily catch those pirates."

If we needed any further proof you don’t understand the concept of "scale". By the time Youtube has enough manual checkers in place it will be employing a sizeable proportion of the US citizenry.

Since you’ll need coverage for about 200 languages on top of that I invite you to reconsider your proposal that the default position of a user content platform should spend hundreds or thousands of times their net revenue on moderation.

Add Your Comment

Your email address will not be published. Required fields are marked *

Have a Techdirt Account? Sign in now. Want one? Register here

Comment Options:

Make this the or (get credits or sign in to see balance) what's this?

What's this?

Techdirt community members with Techdirt Credits can spotlight a comment as either the "First Word" or "Last Word" on a particular comment thread. Credits can be purchased at the Techdirt Insider Shop »

Follow Techdirt

Techdirt Daily Newsletter

Ctrl-Alt-Speech

A weekly news podcast from
Mike Masnick & Ben Whitelaw

Subscribe now to Ctrl-Alt-Speech »
Techdirt Deals
Techdirt Insider Discord
The latest chatter on the Techdirt Insider Discord channel...
Loading...