The European Venue for Green Ideas
Follow us on
Society, Media and Culture

Link Taxes and Upload Filters Will Not Fix the Internet

In September the European Parliament voted in favour controversial online copyright proposals that could see link taxes and upload filters introduced if the legislation passes the final vote after negotiations end in early 2019. We spoke to Pirate MEP Julia Reda on the risks these proposals pose for users, publishers, and European platforms and asked her about the way forward for digital politics in Europe.

Adam Ostolski: What is the link tax? If the proposed directive passes, should I expect a letter from the tax office next time I post a link on Facebook?

Julia Reda: The proposed link tax, which is called the neighbouring right by its supporters, would apply to links or small parts of press articles that are shared on websites and online services. Publishers hope to force companies like Google or Facebook to pay for licences to be able to use their content. The only exclusion would be for purely private activities such as sending a link in an email. But if you post a link on Facebook or edit Wikipedia and add a reference to a news article, then you will need permission, not from the tax office, but from the publisher.

Publishers are some of the biggest advocates of the link tax. Why do they see the need for one?

Publishers claim that links to their news articles on social media are somehow harmful to them and want to force large platforms to pay for them. In reality, publishers all employ experts to make sure that their articles feature prominently and look as attractive as possible on social media and search engines. The commercial activities of the publishers themselves indicate that they benefit from social media linking.

But publishers are certainly having a difficult time enforcing the copyright of news articles. The proponents of the neighbouring right say that platforms use press articles for free and that there is no way to stop them from doing so. But copyright already applies to news articles so platforms cannot simply take a newspaper article and republish it without a license. If they are doing this and the publishers cannot stop them, then the problem is with the enforcement of the law.

What are some of the risks to freedom of information posed by the link tax?

Publishers hope that the proposal will work like a tax. But if you tell a company that they have to pay to link to your site, then they will probably stop doing so. The danger for internet users is that large platforms will decide not to pay publishers and will just block links to news websites. In the future, it could be that when you try to post a link on Facebook, you get a message saying “We’re sorry, linking to this news website is not possible in Europe”. Even if you do not share links often, you will see fewer links to news articles from a less diverse array of sources.

Unlike what many people think, big technology companies are not particularly worried about the link tax. If you compare how much the technology companies have lobbied on taxation or competition, they are not really that active on the proposal. Facebook and Google know that they will not be affected as badly as their smaller competition.

On the side of those opposing the directive we find the Wikimedia Foundation, the parent organisation of Wikipedia. Why are they concerned?

Wikipedia is one of the online services that would be highly affected by a link tax. Wikipedia would have to remove links to news sources from their pages, which would mean that users would access less well-referenced information.

Despite the flaws, lobbying groups like European Publishers Council remain vocal supporters. Are there experiences of countries that have put in place similar charges that we can learn from?

We tried a link tax in Germany and, as soon as the law was introduced, Google simply stopped showing snippets from news articles on Google News and only displayed the headlines, since the German link tax still permitted very short snippets of text. As a result, press publishers soon found that far fewer people were clicking on their links and their visitor numbers dropped. In the end, Google was given a free licence to continue using their snippets while smaller online services had to pay.

Exactly the same thing will happen on the European level. A law originally intended to take money from Google to give to publishers will not create new revenue and instead will further entrench Google’s advantage.

Back in July, when the directive was discussed in the Legal Affairs Committee of the European Parliament, you advocated a compromise. What was that proposed compromise?

Our alternative is the so-called presumption rule. It too would ensure that publishers get paid whenever platforms use whole news articles or significant parts of then. But we proposed that using the headline or any other part of a news article that is not protected by copyright still be allowed. The proposal we put forward says that unless a platform can prove that it has received a license directly from the journalist who wrote the piece, the publisher can enforce the rights on behalf of the author. Proving the exclusive right to represent every journalist who writes for them can be pretty tedious for publishers.

The other controversial part of the legislation concerns upload filters. The proposed directive would make platforms such as Youtube install tools to make sure no copyrighted content is uploaded. How do upload filters work?

Upload filters use algorithms to check uploaded content against a database of works identified to them by rights holders. If uploaded material is similar to a work registered in their database, the platforms deny the upload.

What upload filters cannot do is interpret the context in which a work is used. If you quote a work, which is legal in all European countries to some extent, an algorithm cannot distinguish a copyright infringement from a legal quotation. It’s even more difficult with the right to parody, which exists in some European countries. For an algorithm to be able to understand that an upload is legal, it would have to understand humour. That’s very far from what computers are able to do today.

Instead, what upload filters do is simply delete everything that is a copyright exception, and often works in the public domain too. Ulrich Kaiser, a German music professor uploaded public domain recordings of old classical music. The recordings were immediately blocked by YouTube’s content identification system despite being legal, probably because some copyrighted performances of the same music that sound very similar exist.

So to what extent are effective upload filters even viable technology?

I think the European Commission is trying to legislate for technology that does not exist. Simply stating that platforms should use technologies that can delete illegal content and keep legal content online does not mean that the technology actually exists. Platforms will of course use the kinds of filters that already exist but the fact is that the technology is not that advanced. There is a blind belief that algorithms can solve complex legal questions.

What would it mean to hand over the enforcement of this area of copyright law to algorithms?

If we accept that the copyright law is enforced by automatic algorithms, we are effectively getting rid of the exceptions protecting freedom of speech and putting an end to the public domain. Only the parts of copyright law that algorithms are able to enforce will be enforced, the remaining parts that are supposed to protect users will not.

For smaller European platforms, the responsibility to apply algorithmic upload filters is also problematic. Even if small platforms use the best filters on the market, the filters will still make mistakes. They both delete too much, as with public domain content, but also delete too little, and will not find every single copyright infringement. Platforms will be liable for mistakes made by their filter, creating a strong incentive for platforms to delete more than they have to. One result could be that they only permit uploads of licensed material, which would be a huge problem for fan culture and user-generated content.

What would it mean for big platforms, the target of the legislation?

If you look at the text adopted in the Council, it says that the platforms have to pay the right holders, unless they use filters. YouTube and Facebook are the only companies in the position to use those filters and they could well start selling them to other smaller European platforms. A proposal supposed to be bad for large companies may end up giving them a new business model.

Have you been able to find a way forward or compromise on upload filters?

In the Green Group, we recognise that for most creators the priority is not to prevent people from seeing their work; the priority is fair payment. The solution that we have found is not something radical written by Greens or Pirates, this compromise proposal comes from a Conservative MEP Michal Boni from Poland and was also adopted by the European Parliament’s Internal Market Committee. Unfortunately, this approach so far did not find a majority in the European Parliament.

The compromise says that active platforms that take information users upload and monetise it through targeted advertisement (or similar ways) cannot claim to be neutral and must obtain a license. YouTube, for example, monetises uploaded music. YouTube analyses uploaded tracks, puts it them in sequences, and plays them alongside targeted advertisement, so they cannot claim not to know that people were uploading it.

However, the proposal also states that no platform can be forced to introduce a filter. A platform can be forced to pay under certain circumstances, but it cannot be forced to filter because the European Court of Justice has said several times that filtering obligations violate fundamental rights. They violate the fundamental rights of users to expression and privacy  and the fundamental right of the platform operator running a business.

Who are the allies and opponents of digital freedoms nowadays?

I don’t think that everyone who supports upload filters or the link tax wants to destroy online freedom. Members of the European Parliament are under significant pressure from media companies to adopt these proposals. Journalists and publishers have campaigned on the basis that if the law is not passed how they want, then the news media and quality journalism in Europe will disappear. Not everybody is aware of how disastrous the link tax was in Germany and Spain.

In the European Parliament, the European People’s Party is very supportive of these proposals, whereas the Greens and the Left Group are against and want to better balance copyright and fundamental rights. The Social Democrats are split, so are the Liberals and the Eurosceptics.

It’s not a classical left/right issue. From a generational perspective, younger people who have grown up with the internet are more likely to be critical of the proposals.

Privacy advocates such as the late Caspar Bowden revealed the impact of the US surveillance programmes on EU citizens’ fundamental rights. Some people suggest direct internet connections between Europe and Latin America, bypassing the United States, or relocating some hardware production to Europe. Does ‘digital sovereignty’ offer a way to preserve digital freedoms in Europe?

There is a big difference between regulating the internet in the EU, which is important, and shutting ourselves off from the rest of the world. European internet regulation, such as net neutrality or data protection, could set the standards for the rest of the world. That’s why passing laws together as the European Union is crucial. On copyright so far the EU has failed to provide a common answer. Even under the proposed directive, there would still be completely different rules on what is considered, for example, a parody or a quotation in every European country. It attempts harmonisation for rights holders and imposes new obligations such as filtering from the European level, but it does not offer the same protection for users.

Today many internet companies are not enforcing European copyright laws at all. Have you tried to go to YouTube and say, “I’m sorry, you took down my video, but actually it is allowed under the Polish exception for education”? YouTube does not care about Polish or German copyright law. They enforce US copyright law and the only way for us to have more influence over the rules is make them together.

Take the case of the data protection regulation (GDPR) that came into force in May 2018. Even though it is not perfect, most American companies, and eventually most Chinese companies, will try to comply with GDPR. The European market is too big to ignore. The same could happen for net neutrality.

But we should be very careful of balkanising the internet with walls between the European internet and the rest of the world’s internet. The value of the internet is that it is global. As Pirates and Greens, we should be globalist and make sure that the global rules exist to respect fundamental rights not just to protect us from outside influences.

You wouldn’t be enthusiastic about the idea of a European cloud?

There is nothing wrong with a European cloud; the question is what it is for and whether you can communicate with other places. Having a European cloud and the free flow of data within the EU is fine. But the rules we set for the internet should have the ambition of encouraging other countries to follow. The EU should also look to the kinds of digital rights being developed in other parts of the world such as India or Mexico.

The movement for digital citizenship, including Pirate parties, has been more successful in stopping bad legislation than in advancing its own goals. We stopped the Patents Directive in 2005. We stopped the Anti-Counterfeiting Trade Agreement in 2012. The copyright directive has the final vote in Parliament early next year, so can still be stopped. Regardless of the outcome of this fight, what comes next?

Civil liberties organisations and fundamental rights movements lack the funding enjoyed by large companies. We are on the defensive and don’t have the same long-established relationships with the most powerful layers of politics as entertainment companies.

The best way we can bring about change is to make a cultural shift and convince people that being open and sharing is beneficial to them. The amount of money that people spend on culture has not decreased. In the music industry, people spend less on records and more on concerts, from which musicians get a bigger share of proceeds.

Many creators today are building their careers on the principle of openness, using Creative Commons or other means of free sharing. And people are still willing to support them and to pay. We need to acknowledge that people want to experience culture and want to support creators, and give them opportunities to do so.

What do Pirate parties, and the Pirate project, have to offer Europe?

Cultural shift is at the heart of the Pirate project. In a similar way to the Social Democrats and the industrial revolution or Green parties and climate change, Pirates are interested in the social changes associated with the digital revolution. Our goal as Pirates is not to turn back the clock, as some established parties would like. We don’t vilify digital technology but see it as an opportunity, recognising that we need new social institutions so that digital technologies work for the benefit of the people and not just as tools for surveillance.

To do that we emphasise the commons. We support things like free and open-source software and community-owned networks. Our answers for technology are quite similar to the answers the Greens are finding for energy. We have energy cooperatives and in the same way we need digital infrastructure cooperatives to make sure that digital technology is in the hands of the people, not multinationals and autocratic governments.

This is the Pirate vision for the future: digital technology controlled by the people.

Newsletter

Link Taxes and Upload Filters Will Not Fix the Internet

Cookies on our website allow us to deliver better content by enhancing our understanding of what pages are visited. Data from cookies is stored anonymously and is never shared with third parties.

Find out more about our use of cookies in our privacy policy.