CMU Trends

Trends: Safe Harbour In Five Steps

By | Published on Monday 3 September 2018


Safe harbour reform is big news in the music community again as the European Parliament prepares for another debate on the proposed new European Copyright Directive.

Article thirteen of that directive seeks to reform the so called safe harbour that protects internet companies from liability when their customers use their services to infringe copyright. However, said reform has proven controversial and was a key reason why the Parliament previously voted down the most recent draft of the copyright proposals. Keen to ensure it gets passed next time round, the music industry has recently ramped up its long-running ‘value gap’ campaign.

But what is the copyright safe harbour? What specifically does the music industry have against it? And what does that value gap campaign hope to achieve? CMU Trends explains it all in five steps.

To understand safe harbour we need to say a few things about copyright and copyright infringement. Copyright protects a variety of different kinds of creative work – or content if you prefer – and in the music industry we are most interested in the separate copyrights in songs and sound recordings.

Copyright law provides the copyright owner with a number of controls over their content. This means they can control what happens to their song or recording, or whatever. The exact list of controls, and the terminology used, varies from country to country, though UK law identifies six main controls: the reproduction control, the distribution control, the rental control, the adaptation control, the performance control and the communication control.

Because of the controls, only the copyright owner has an automatic right to reproduce, distribute, rent out, adapt, perform or communicate the song or recordings that they own.

If anyone else wants to do any of those things: if they want to make copies of a song or a recording; if they want to sell or rent out those copies; if they want to adapt an existing work; if they want to perform a song or recording in public; or communicate or broadcast a track to the public; then they need the copyright owner’s permission.

The copyright owner will sell this permission, which is how copyright makes money. This selling of permission to third parties is called licensing.

If a third party exploits one of the controls of the copyright without permission – so maybe they make a copy of a song or recording without permission, or they adapt a track without permission, or they perform a song in public without permission, and so on – that is copyright infringement. When someone infringes someone else’s copyright, the law says that the copyright owner should sue the infringer for damages.

Quite how this works varies from country to country. On a basic level it goes like this: I own the copyright in a song or recording, which means I can control what happens to that song or recording. If you exploit one of my controls without permission then the law says I should sue you. So, in the case of online infringement, if you share a track that I control via the internet without my permission, then the law says I should sue you for damages. Simple.

However, in order to share my track online you probably utilise the services or the technology of one or more internet providers. For example, you will probably use an internet service provider to connect to the internet in the first place. You might then store your unlicensed copy of my recording on somebody else’s server or in a digital locker. Or you might use a social media platform or user-upload platform to share the track. Let’s call all those companies the ‘intermediaries’.

As a copyright owner it would be much easier for me to sue the intermediary rather than suing you directly, for two key reasons. First, the intermediaries are much easier to find. Secondly, they probably have more money. There’s no point suing people or companies who don’t have much money, because even if you win they can’t afford to pay you damages.

Copyright law often allows you to sue individuals or companies who help other people to infringe copyright, even if they are not directly involved in the infringement. Quite how this works varies greatly from country to country, as does terminology used, with this kind of infringement being variously known as ‘secondary infringement’, ‘contributory infringement’ or – as in the UK – ‘authorising infringement’.

However, when it comes to our online example, where you’re sharing a copy of my track without licence via an internet company, the chances are that I can’t sue the intermediary because of this thing called ‘safe harbour’.

The safe harbours were added to copyright law back in the 1990s when mainstream adoption of the internet was first starting to take off. Internet companies which were starting to provide their services to a much wider range of customers – including individuals rather than just other businesses – pointed out that if they could be held liable for the copyright infringement of every single person using their networks and servers, then that liability would simply be too high, and that would make it impossible to make internet access available to all.

The safe harbour aimed to tackle this problem. It was decided that the provider of internet services could not be held financially or criminally liable when their networks or servers were used by third parties to infringe copyright. But on one condition. To qualify for safe harbour protection, the intermediary must operate what is often referred to as a ‘takedown system’ to help rights owners deal with infringers and infringing content on the intermediary’s network.

This means that if a copyright owner sees that its content is being infringed by an intermediary’s customer, it can alert that intermediary to the infringement, and the intermediary is then obliged to stop the infringement, which usually means deleting the infringing content from their network or server. Providing that an internet company that is acting as an intermediary in the unlicensed delivery of my content offers a takedown system of this kind, they enjoy safe harbour protection, which means that I can’t sue them for copyright infringement.

Quite how the safe harbour works varies around the world. On a global level, most attention has traditionally been given to the safe harbour under US copyright law, which originates in the Digital Millennium Copyright Act. Although more recently a lot of the conversation has actually been about the safe harbour under European law, which originates in a European Union e-commerce directive from 2000.

The music industry doesn’t actually have a problem with the general principle of safe harbour, but it does have some issues with the way that it has been implemented over the last two decades.

The first key issue relates to the obligations of companies claiming safe harbour protection, and in particular the obligation to operate a takedown system. The law is often somewhat vague on what a takedown system should actually look like. Copyright owners and safe harbour dwellers have therefore tended to look for guidance on the exact obligations of the latter whenever landmark copyright cases have gone to court.

In the early days of digital, copyright owners often felt that judges tended to set the bar too low when it came to these obligations. Because of this, some internet companies claiming safe harbour protection were able to operate pretty mediocre takedown systems. Copyright owners suspected that certain internet companies were operating deliberately mediocre takedown systems, so that they could claim safe harbour protection without actually removing much infringing content.

Why would they want the copyright infringement to continue? Possibly because they have a business that actually relies on having access to a large quantity of unlicensed content. Or simply because they don’t want to annoy their copyright infringing customers by blocking content, sending out stern warning emails and disconnecting the accounts of repeat infringers.

Many of the high profile legal battles that have seen internet services accused of operating deliberately mediocre takedown systems have taken place in the US. Both streaming service Grooveshark and digital locker MegaUpload were accused of operating deliberately mediocre takedown systems because their businesses relied – the entertainment industries argued – on having a steady stream of copyright infringing content being uploaded by their users.

Though in the Grooveshark case, the music industry was actually nervous about testing its argument in court, as it wasn’t clear whether or not it would win. So instead it employed a number of legal technicalities to justify litigation, including the allegation that staff-members as well as users had uploaded unlicensed music. If Grooveshark employees were directly involved in the infringement, the safe harbour would not apply.

The MegaUpload case resulted in both criminal and civil proceedings. The former have got caught up in years of legal wrangling in New Zealand, where various MegaUpload executives are fighting extradition to the US to face the criminal copyright charges. Meanwhile the civil case is on hold pending the outcome of the criminal proceedings. All of which means the entertainment industry’s allegation that MegaUpload claimed safe harbour protection while only paying lip service to its takedown obligations is yet to be properly test in court.

The other big cases on this point in recent years have involved American internet service providers Cox Communications and Grande Communications, both of which have been accused of not fulfilling their obligations under the aforementioned DMCA in order to qualify for safe harbour protection.

BMG sued Cox making this very argument and won at first instance, with the court deciding Cox was not eligible for safe harbour protection because of its deliberately mediocre system for dealing with repeat infringers. The case was overturned on appeal, though mainly due to technicalities, with the appeals court judgement pretty much endorsing BMG’s core argument. Cox then settled with BMG.

The Recording Industry Association Of America sued Grande and then, more recently, Cox. The RIAA v Grande case continues to go through the motions and further tests the obligations of safe harbour dwellers. The music industry hopes that the precedent set in Cox – especially if it is endorsed in Grande – is slowly increasing the obligations and liabilities of safe harbour dwelling internet companies – especially ISPs – under American law.

Away from America’s courtrooms, many rights owners argue that even good takedown systems put too much strain onto the copyright owner, who is forced to constantly monitor the networks of countless intermediaries looking for infringing content, and to then issue a steady stream of individual takedown notices requesting that that content be removed. Technologies have been developed to help with this process of course, but it is still a major task for copyright owners to undertake on a daily basis.

This strain is all the more severe because as soon as a piece of content is removed by a takedown notice, what frequently happens is that the same content is re-uploaded to the same network or server. So although in theory the content has been removed, it is still available – without licence – elsewhere on the intermediary’s platform.

It’s with this in mind that the issuing of takedowns under the safe harbour laws is often compared to an endless game of whac-a-mole. As soon as you have one bit of content taken down, the same bit of content pops up somewhere else on the same platform.

Many rights owners insist that the intermediaries should play a much more proactive role in keeping infringing content off their networks. And in particular, that they should develop what is often referred to as a ‘takedown-and-stay-down’ system. So that once a piece of content has been removed off any one intermediary platform, if another user attempts to upload the same piece of content without licence, the intermediary should automatically block the content on the second, third, fourth, fifth – and so on – upload.

Some internet companies have invested in developing technology that, in theory at least, does automatically block content that has previously been removed if and when it is reuploaded. Although music rights owners would like the employment and enhancement of such technology to be a prerequisite of safe harbour protection.

The second key issue is the one that has got the most attention in recent years and is where the music industry goes into battle with Google’s YouTube. Despite it, somewhat ironically, having one of the better takedown-and-stay-down systems in the form of its Content ID technology. The issue here is the range of services claiming safe harbour protection. Which is to say, what do we mean by ‘intermediary’?

In recent years the music industry has argued that user-upload platforms like YouTube shouldn’t enjoy safe harbour protection, however good their takedown systems may or may not be. Moreover, the argument goes, law-makers never anticipated YouTube-type services enjoying safe-harbour protection when they came up with the concept in the 1990s.

Many user-upload platforms provide content-sharing tools for content creators – which might be classified as ‘intermediary’ services – but then they aggregate and redistribute that content through a home page, search engine and playlists. In doing so they create their own content platform that they monetise through advertising. That, the music industry reckons, takes YouTube type platforms beyond being a mere intermediary and therefore beyond safe harbour protection.

If YouTube lost its safe harbour protection, then copyright owners could sue Google whenever its users upload copyright infringing material. That would force YouTube to take responsibility for the content its users uploaded, or at least the content it then aggregates and redistributes through its home page, search engine and playlists.

That said, nobody actually wants to sue YouTube. The issue here is that the existence of the safe harbour greatly weakens the music industry’s negotiating hand when talking to the Google service about licensing deals. To understand this argument, we need to consider how digital licensing normally works.

If you are a streaming service like Spotify or Apple Music, you need to get permission from the record companies, distributors, music publishers and collecting societies to exploit the songs and recordings that they respectively represent.

In order to operate your streaming service, you need permission from all of these copyright owners to exploit the relevant controls of their copyrights. So you sit down with each record company, and distributor, and music publisher and collecting society, to agree terms which will be set out in a licensing deal.

As a streaming service, you have to do these deals in order to access the content. This is particularly true with your record company and distributor deals, because at the end of the negotiations, those companies will actually pump recordings into your platform.

If you don’t agree a deal, then you don’t get the content. The music companies can exploit that fact to their advantage. The music companies want to do a deal, because they want to make money out of their recordings or their songs, and we all know that streaming is now the key revenue stream in the recorded music sector.

However, while the music companies want to do deals with streaming services, they can still exploit the fact that – if no deal is done – the service doesn’t get the content that it needs to operate. That negotiating power enables them to push up the price of accessing their music.

This is how licensing negotiations occur between the music companies and the classic streaming services like Spotify and Apple Music, and most of their direct competitors. But with user-upload services, like YouTube, the service often already has most of the content on its platform, because its users have already uploaded it.

More importantly, if no deal is done, rather than the user-upload service having no content – because it still has the content its users uploaded – instead the music company must now request that all that user-uploaded content be removed through the intermediary’s takedown system. So, in the case of YouTube, that means removing each song and each recording through the Content ID system on day one, and then ensuring that content stays down.

As we have already said, Content ID is a pretty good takedown system, but it’s not perfect, especially on the songs side. So music owners need to keep monitoring the platform to ensure that its songs and recordings have not resurfaced, requesting new takedowns whenever they do. And because of safe harbour, it’s for the label, or distributor, or publisher, or collecting society, to pay for the people required to manage this takedown process, both initially, and on an ongoing basis.

With all that in mind, it is simpler for the record company, or distributor, or publisher, or society, to just do a deal with YouTube – even an unfavourable deal with YouTube – because at least then, when their content is uploaded to those user-upload platforms, they will earn some royalties. Which can pay for the cost of administrating a Content ID-type platform and hopefully provide a little bit of profit margin.

YouTube can exploit this fact to its advantage at the negotiating table. It can pressure the music companies into the kind of deals that they would never agree to with the other streaming services like Spotify and Apple Music.

The music industry argues that this is exactly what has happened as it has agreed licensing deals with YouTube. So while the Google service is licensed by most labels, distributors, publishers and collecting societies, it enjoys much better deals than the audio streaming services like Spotify and Apple Music. Even though, as the digital market has shifted from downloads to streams, YouTube has become a head-on competitor to the digital music companies the music industry relies on for much of its revenues.

The difference between the monies YouTube pays to the music industry under its more preferential deals and the monies services like Spotify and Apple Music pay is what is often referred to as the ‘value gap’.

So, what can be done about all of this? Well, the music industry wants safe harbour laws to be rewritten – or at least clarified – so that user-upload platforms like YouTube no longer qualify for protection. The result of that, they feel, would be that their negotiating hand would be greatly strengthened when they sat down with companies like YouTube to negotiate new licensing deals. Which means that YouTube would be forced to pay rates more in line those paid by Spotify and Apple Music. Needless to say, the user-upload platforms object to that proposal.

In the US, the country’s Copyright Office instigated a review of the DMCA safe harbour at the end of 2015. As a result, the American music industry began campaigning more vocally against safe harbour in general and YouTube in particular. The Copyright Office is yet to feedback on that review, though we are expecting some kind of report later this year. Even if that report supports safe harbour reform, it seems unlikely any major legislative proposals to that effect will go before Congress in Washington anytime soon.

In Europe, a major review of copyright law was instigated in 2015 as part of the European Union’s Digital Single Market initiative. The music industry successfully managed to use that review of copyright law to get safe harbour reform into a draft new copyright directive that is currently working its way through the law-making process.

An initial draft of that directive was published in September 2016 and article thirteen deals with safe harbour. It doesn’t outright deny safe harbour protection to user-upload platforms, but instead seeks to increase the obligations of platforms of this kind in order to enjoy said protection.

As is often the way with new legislation, there were some ambiguities in the original draft that could possibly allow some user-upload platforms, including YouTube, to insist that they were already compliant with the new demands. Although the music industry responded more positively to the proposals than the tech sector, which suggested that the Googles of the world were nervous that their liabilities would indeed be increased. Either way, both sides have been lobbying hard ever since to try to get article thirteen amended to their advantage.

The most recent draft presented to the European Parliament was generally seen to favour the music industry, although it was still arguably open to interpretation. Clearly the tech sector is worried because, after that draft was approved by the Parliament’s legal committee, lobbyists for big tech went into overdrive. As a result, MEPs at large voted down the current version in July, with further discussion now scheduled to take place on 12 Sep.

Article thirteen has proven to be one of the most controversial elements of the new copyright directive. Critics say the proposed new law will have unintended consequences that will negatively impact on everyday internet usage, on the sharing of ideas and memes over the net, and of the development of new online businesses.

Lobbyists for the music and wider media and entertainment industries dispute these claims. They have also accused the tech sector, and the copyright critics they support, of spreading misinformation and utilising online lobbying tools to make it appear like opposition to copyright reform is much more widespread than it really is.

A lot of the campaigning against article thirteen has centred on the potential impact it could have on individual web users sharing content through platforms like YouTube.

That campaigning often ignores the fact that, as an individual web user, you are already infringing copyright if you use someone else’s music or photography or film – or whatever – within any content you upload. And you could in theory be sued for that infringement.

The music industry’s argument is that the intermediary, which is already monetising the uploaded content, should be obliged to agree licensing deals with the copyright owners. Doing so would protect the intermediary from any new liabilities under article thirteen, but also the user from their liabilities under existing copyright law.

Critics would likely counter that the uploader is, in fact, allowed to use copyright material without licence in certain scenarios as a result of the so called ‘copyright exceptions’. So this would include whenever copyright material is used for the purposes of critical analysis or parody. Could the content filtering systems that article thirteen might require services to implement be capable of spotting this kind of ‘fair use’ or ‘fair dealing’ of copyright materials? Probably not.

This challenge was actually discussed – although not resolved – in the famous ‘Dancing Baby’ case in the US, which considered the responsibilities of copyright owners utilising takedown systems, especially when the use of copyright material was fair use under American law. You also have the added issue that – even if an AI technology could be trained to spot fair use – copyright exceptions vary from country to country. So it might be hard to ensure that any increased content filtering as a result of article thirteen wasn’t inadvertently blocking fair use material.

The music industry would argue that navigating copyright exceptions is a challenge to be met by the technology companies and the copyright owners, and not a reason to allow user-upload platforms to enter the music or video on-demand markets without securing the required opt-in licences or paying the market rate royalties.

It remains to be seen what happens in the European Parliament this month. And that’s not the end of the debate. European Union directives must be approved by both the Parliament and the EU Council, the latter of which brings together the most relevant minister from each of the 28 member states.

Once passed, it must then be implemented at a national level by each member state. If the likes of YouTube decide to dispute whether or not the new law actually increases their liabilities, the whole thing will have to be tested in the courts. And not the English courts, because by that point the UK will probably have exited the EU.

Alongside all this, YouTube is trying to placate its critics in the music industry by pushing its own premium streaming service licensed more in line with Spotify and Apple. If YouTube Music can gain momentum, then Google can try to justify the lower royalties paid on the main YouTube platform by arguing the free service helps upsell the paid-or service.

Meanwhile, if the streaming sector at large continues to grow at its current pace – and if the streaming services themselves can become profitable – perhaps an increasingly affluent music industry will ultimately learn to live with a few user-upload platforms exploiting a safe harbour loophole to lower the royalties they pay whenever music is streamed.

All of which means that – while this month will be a key chapter in the story of the music industry’s battle with safe harbour and YouTube – it is unlikely to be anything like the final chapter.