On 30 November, Eric Luth at Wikimedia Sverige was invited to give a speech on Wikimedia, Wikipedia, and copyright at an interdisciplinary IP symposium in Stockholm. This is his speech.
In this talk, I’ll try to give a few of my perspectives on the internet legislation that shapes our work, dive deeper into copyright, and then give a few of my thoughts on where copyright legislation is lacking.
But to begin with, just a few words on the Wikimedia movement.
We are the movement behind platforms such as Wikipedia, Wikimedia Commons, and Wikidata.
Wikipedia alone today contains more than 50 million articles in over 300 languages, read on average more than 250 million times a day. Wikimedia Commons, the sister project for images and free media files, just passed 100 million uploaded files. All created by users, for users, and everything is published under free licenses, so-called Creative Commons licenses.
The reach on the platforms is astonishing. But still, that is only the direct reach. When we start to consider the indirect reach, through, for example, ChatGPT, Siri, or Alexa, the numbers – though harder to measure – get even more astonishing.
Two years ago, we celebrated Wikipedia’s 20th birthday. Despite having radically increased access to knowledge worldwide, and created one of the largest compilations of knowledge in human history, Wikipedia is still, I think younger than all of us in this room.
Young, but in an Internet context extremely old. Wikipedia was created on an entirely different Internet, and few, close to no, other of the other websites created back then, are still around today.
This was an era of experimentation and creativity. People, mostly individuals, sat at home, I imagine in their basements, trying out things. Some things, like Wikipedia, worked, some things didn’t. It was a playground for new ideas. For most, money was not the driving factor, ideas were. Most websites were run non-profit, simply because there was no profit to make.
The situation has changed dramatically since. Depending on how you define it, Wikipedia is the only, or one of very few, non-profit websites among the 100 most visited websites in the world. The commercialization of the Internet has been quick and has swept away most of the non-profit playground experimentation, in exchange for large, global, commercial super platforms.
The start of this change dates back to the middle of the first decade of the 21st century when a few actors realized the commercial potential of user data, a potential that was realized during the last decade to a level I think none or few of us were able to predict.
My point is not to say that this change is inherently good or bad, but it is a fact that the Internet today is a different place than it was when Wikipedia was born.
To a certain extent, this is the result of these large commercial platforms, run by creative entrepreneurs who started with playgrounds but created new worlds.
But what I would like to argue is that this is also the result of the legislative development. Politicians and lawmakers have, for ideological or practical reasons, tried to curtail the large commercial platforms, but the effect in most cases has been that it is only those platforms that can afford to pay for the solutions required.
So while lawmakers have had the ambition to challenge the close-to-monopolistic situation of the large commercial platforms, in many cases they have rather reinforced this semi-monopolistic situation – by forcing solutions that are so expensive to implement, that smaller or non-profit actors cannot do it.
I think this is especially the case in Europe, where European politicians ponder why there are no large European platforms while developing legislation that effectively stifles innovation.
I understand the wish to legislate the internet from lawmakers, and get me right: legislation is needed. There are indeed actors that have actively aspired to share and distribute copyrighted works illegally. It would be hard not to mention the Pirate Bay in this context.
We also see an increasingly difficult online environment, where the tools that mean so much freedom for so many of us can also be used for criminal purposes. Online platforms need to take responsibility for the content on their platforms, I think it’s a given, and also take a social responsibility that goes beyond their own platforms in themselves.
But most serious platforms would rather see each case of copyright infringement as a failure of the model of the platform, pointing to a need to refine the model and learn from the mistake.
One of the most important pieces of European legislation for online platforms has for years been the eCommerce directive, which establishes a so-called notice and take-down model – similar to the Digital Millennium Copyright Act in the United States.
Under this legislation, Wikipedia and the other Wikimedia platforms would need to analyze and take action as soon as a notice is received from a rightsholder.
Under this model, of course, there might be a few works – among the, for example, 100 million files on Wikimedia Commons – that are infringing copyright. But as far as we know, the number is very small.
Every month, the community acts upon ~600 requests to delete content because of copyright infringement. Requests that have been made by other volunteers, that is, that are keenly reviewing all change logs to detect potentially infringing material.
Yesterday saw 100 requests from users about files where copyright was unclear. For example a set of images from the university library of Basel. The Freedom of Panorama of Switzerland only covers the exteriors. There was a renovation in the 1990s, but it is unclear to what extent these stairs were renovated in the 1990s, and if the architect of the renovation retains any copyright beyond the original architect from the 19th century. This is the kind of case that is always discussed, openly and transparently. We still don’t know if these stairs will survive online.
During the first half of 2023, rightsholders, on the other hand, gave less than 30 notices to the Wikimedia Foundation, of which a few were granted under the Digital Millenium Copyright Act. Many of these cases are indeed difficult grey zones and are reviewed very carefully by professional lawyers and cultural heritage practitioners.
To make this very clear: 600 cases are opened by the community itself every month. The first half of 2023 saw 30 cases, altogether, based on notices from rightsholders. The community moderation works, in most cases, very well.
But back to the legislation. In the early days of the Internet, this notice and take down approach was the standard or at least very common procedure, as outlined in the EU eCommerce directive as well as the named US Digital Millenium Copyright Act.
But as you may know, this has changed profoundly in the EU during the last few years, starting with the Copyright (CDSM) Directive and followed by the Digital Services Act. There is an increasing trend to require proactive work to make sure that infringing or unlawful content may not be uploaded at all, rather than being deleted upon request. Increasingly it requires proactive work to make sure that deleted works are not uploaded again, in what has been called “notice and staydown”, which is the foundation of Article 17 of the EU Copyright Directive and indirectly, though perhaps not directly, requires upload filters. Despite the prohibition against general monitoring.
I find it hard to see how this piece of legislation can benefit anyone else than the large commercial platforms.
Thankfully, Wikipedia is exempted from this law, though I must say that the article that explicitly exempts Wikipedia (or rather “not-for-profit online encyclopedias”), Github, and others is a very bad way of writing law
But we are not exempted from similar provisions in the DSA. It will be interesting to follow the development ahead. The European Commission has listed 17 platforms that are considered very large online platforms. Of those 17, Wikipedia is the only non-profit. But what it means is still unclear. Wikipedia works in 318 languages. Are all languages covered, or only English? Only EU languages? Who decides and who supervises these questions? As I assume that you can understand, having volunteers monitor 50 million articles in 318 languages is not a piecemeal task.
I hope we can survive this legislation, and show that you can be a very large online platform, run as a non-profit, on the internet, and still survive in 2024. Show that a volunteer community can moderate content, and not only employed staff.
But honestly, I don’t think that this is important only for us. If the EU would kill the only non-profit platform by its laws, I think it would be a large shame for the EU as well.
The two decades of community moderation has had another effect. I usually say that the active users on Wikipedia, Wikimedia Commons and other Wikimedia platforms are among the most copyright-informed people I have ever come across. Few have a legal background, but they read the law, try to understand it, and apply it – according to their interpretations – in thousands and thousands of cases.
The work of these volunteers shows the importance of copyright legislation that is understandable by laymen and not only professional lawyers.
The discussions around the copyright status of things as graffiti, toys, and model cities could be published as books. I’m constantly fascinated by the depth and width of arguments brought up in discussions around, for example, the famous Raising the Flag on Iwo Jima, one of the most famous photos in US history. There have been dozens of deletion requests about this photo over the years, and the question is still contentious. In the US, between 1928 and 1963, you had to renew the copyright after 28 years, but no one has managed to find the renewal from the photographer – although Associated Press ascertains that the copyright was renewed.
Another contentious topic is which logos of institutions actually reach the threshold of originality and not. Such discussions cover every potential aspect of the questions, one argument more creative than the other.
The volunteers read the law, systematize it, compare it, try to understand it. Help pages on Wikimedia Commons summarize copyright legislation per country – ideally all countries across the globe – around derivative works, freedom of panorama, threshold of originality, de minimis, stamps, you name it.
I don’t want to say that all of the interpretations are correct. I am not a lawyer, but with the basic legal knowledge I have after years of working with copyright advocacy, I know that in fact, many interpretations are not correct. But I would like to argue that, based on the hundreds of pages of discussion about small aspects, these interpretations are most often reasonable interpretations. If there is a flaw in the interpretation, that most likely reflects an unclarity in the law.
One area of especially large concern is the cross-border usage of works. Copyright law, in the 21st century, is still extremely national. Two comprehensive copyright directives from the EU – the Infosoc and the CDSM directives – still fail to properly harmonize the copyright laws of the EU. We cannot work as one European movement, one global movement. We have to dive deeply into every one of all jurisdictions of the world, of Europe.
Let me give one example: the Hundertwasser decision from the German Federal Court of Justice from June 2003. The decision concerns the Freedom of Panorama exception, which lets users take photos of work permanently placed in the public space and share these freely, with some possible restrictions. The FoP is one of the exceptions allowed under the so called Infosoc directive, but not mandatory. There are also several discretionary words in the Infosoc wording, which has opened up for different national implementations.
So in the Hundertwasser example: the photo was taken of the Hundertwasser house in Vienna. According to the Austrian FoP version, the building has to be located on a public space, but in the German FoP version, the user has to be on public space. The photo of the building was taken from the roof of another building and across the street. This photo was perfectly legal under Austrian law, whereas a German federal court deemed it unlawful.
It is not enough to know if there is a freedom of panorama exception or not. You have to know the exact preposition of the wording. And you have to remember, in all the countries you visit, if this was the building on public space country or user on public space country.
All national jurisdictions obviously have the freedom to, within the framework set up by the EU, the EU Court of Justice, and the so-called three-step test, implement exceptions and limitations in the Infosoc Directive as they see fit. But every language and country-specific “creative solution” adds an extra burden to the work of users active across borders.
I often notice the tremendous difference between the EU copyright framework and the US copyright framework, where most of the Wikimedia servers are based. Our policy is that, for a work to be uploaded to Wikimedia Commons, it needs to be free in the US and in the country of origin. But whereas the Fair Use doctrine takes care of most of the grey zones in the US, in Europe, we have not only to read the law, but the legal commentaries, recitals, and court cases. Of course, we have studied the law in the US too (such as the famous monkey selfie court case), but to a much less detailed degree than in the EU, with its nitty-gritty provisions and patchworky setup.
Where the Fair Use doctrine changes with time and adapts to a changing environment, the EU’s very narrow version of the three-step test makes copyright a bit more closed for every technical innovation.
My firm belief is that it should be easy to do things right, especially when it comes to exceptions and limitations to copyright. Copyright is strict: in the firmest sense, the rightsholder has the exclusive right to their work for all their lives, and in Sweden, 70 years after. The exceptions and limitations give some flexibility to this exclusive right, to the benefit of freedom of speech and the general public.
This means that exceptions and limitations are fundamental for the rights of users. That also means that users must be able to understand the exceptions and limitations, and know how to apply them. Individual users and volunteers don’t have the means to do thorough legal investigations. But they need to be able to base their work on the copyright law.
What I am aiming at is that the copyright legislation needs to be better harmonized on the EU level and that the exceptions and limitations need to be easier to understand. We need less creative variations between countries. We need predictable and clear laws.
I have a dream, and it is that I one day will live in a Europe with a copyright law that does not lead to 100s of pages of creative interpretations on Wikimedia Commons, and where volunteers can spend their time sharing knowledge rather than deleting it based on difficult legal provisions.
It might make my job a bit more boring, but it will make much more knowledge flourish.