Today Germany finally published its long awaited proposal for implementing Article 17 into national law. Back in April of last year, at the occasion of the final vote for the DSM directive, Germany had issued a statement in which it hat expressed the intention to implement the directive in a way that would avoid upload filters. And while today’s proposal breaks this promise (unsurprisingly there wil be filters) it reflects an honest intention to avoid their use as much as possible.

Diverging from the approach followed by all other Member States so far, Germany is proposing to implement Article 17 in a stand-alone law, the “Gesetz über die urheberrechtliche Verantwortlichkeit von Diensteanbietern für das Teilen von Online-Inhalten” or “UrhDaG” for short.

This proposed new law goes beyond merely transposing Art 17 and adds a number of interesting (and potentially controversial) provisions, which seem to be designed to strengthen the position of both users and individual creators. These include the addition of a remunerated de-minimis exception intended to safeguard common types of so-called “user generated content” (§6), the ability for uploaders to “pre-flag” legitimate uses of protected works in their uploads (§8), and the addition of a direct remuneration right intended to ensure that individual creators benefit from the new legal regime (§7).

Leaving aside any ideological preferences, the most notable thing about the German proposal is that it feels like the first attempt so far to make Article 17 “work”. The Ministry of Justice has clearly put a lot of thought into how the provisions would affect uploaders, creators, other rightholders and service providers. This has resulted in a proposal that imagines Article 17 working in a specific way, which is the complete opposite of the proposals we have seen form other member states that simply implement the often contradictory provisions of the directive and leave it to platforms and rightholders to resolve the mess. If nothing else, the ministry deserves to be applauded for this willingness to make choices.

As Péter Mezei points out on twitter it seems that Hungary is now the 2nd EU member state (after France) that has implemented a part of the DSM directive into national law. Yesterday, the Hungarian parliament passed the law “on transitional arrangements and health preparedness for the cessation of emergencies” which through Articles 323 to 327 modifies the existing exception for educational use to also allow digital uses of works both onsite and through secure electronic environments in line with Article 5 of the DSM directive.

This sudden implementation of Article 5 comes as a bit of a surprise, given that until earlier this month Hungary was still consulting on its national implementation of the DSM directive including the educational exception in Article 5.

As it turns out, Hungary had already made the changes to its educational exception on 15 April by way of an emergency decree signed by Prime Minister Victor Orban as part of a wider set of measures addressing the impact of the Corona pandemic (at the end of March the Hungarian Parliament had, in a widely criticised move granted Victor Orban wide-ranging emergency powers that allowed him to rule without any parliamentary oversight). The new law passed yesterday transforms a large number of these emergency measures (including the modification of the education exception) into permanent law.

What makes this whole situation remarkable (apart from the fact that modifying sustantive copyright law by decree as such is most likely unprecedented in the EU context) is, that on substance, the idea to expand the education exception to facilitate online education under the conditions of widespread lockdowns makes a lot of sense. But it is highly disturbing to realise that the only EU member state who has acted in line with this insight is the one run by an authoritarian government and that the ability to implement this policy is the result of its willingness to completely suspend democratic oversight.

In reaction to the Australian plans to make Google and Facebook pay for featuring new media content in their products, Facebook has now one-upped last years statement from Google in France and told press publishers that they can go screw themselves:

If there were no news content available on Facebook in Australia, we are confident the impact on Facebook’s community metrics and revenues in Australia would not be significant, because news content is highly substitutable and most users do not come to Facebook with the intention of viewing news. But the absence of news on Facebook would mean publishers miss out on the commercial benefits of reaching a wide and diverse audience, and social value would be diminished because news would be harder to access for millions of Australians.

Unfortunately for the publishers this is of course a pretty accurate description of the dynamics at play here. As i have argued before, from the perspective of the platforms it makes no sense whatsoever that they should pay for delivering customers to press publishers. As Facebooks response above shows, the only effect of such proposals is to make the platforms mad, and walk away from the table. This is precisely the reason why, instead of fabulating about questionable value transfers, we should rather think about taxing the value that platforms extract from the system as a whole.

From the Verge:

Twitch plans to start automatically scanning clips of live streams for copyrighted music following a wave of takedown requests on years-old videos, which has frustrated streamers over the last week. Twitch says it will automatically delete clips with copyrighted music in them and that it will not penalize streamers — under its current rules, streamers can get strikes for copyright violations that could ultimately lead to a ban.

These takedown requests come from the RIAA which represents the major record companies that have been pinky swearing in the Commission’s stakeholder dialogue that they are not engaging in mass scale takedowns on YouTube. Yet here they are, going after fairly insignificant uses of music in old twitch clips (which, by definition, are shorter than a minute). Anyone who still believes that large corporate rightholders will not use all technological measures at their disposal to go after unauthorised uses of their works, now matter how insignificant, is a fool.

And the fact that a major platform like Twitch (which is owned by Amazon and thus has near indefinite ressources) can be pressured to delete wholesale all clips that contain copyrighted music (or at least the repertoire represented by the RIAA) does not bode well for what we can expect once Article 17 is in force and all platforms operating in the EU will be required to make use of the “services” offered by Audible Magic and the like.

The Kluwer Copyright Blog is running a series of posts by Remy Chavannes on the proposed Dutch DSM directive implementation law. Based on the first two posts (the third is coming tomorrow) it is a pretty good summary of the more controversial elements of the law that accurately identifies the weaknesses of the proposal.

Recommended reading for anyone interested in understanding the legislative wiggle room and the resulting tradeoffs in implementing the DSM directive. Here is the high level summary:

Both the timing and the content of the legislative proposal show an acute desire to avoid the risk of late or incorrect transposition. In the operative provisions and explanatory memorandum, the government has sought neither to mitigate the directive’s limitations nor to offer original interpretations of its impossible compromises. This deliberate lack of imagination is arguably a quality and, while not without its own risks and drawbacks, is based on a pragmatic assessment of national legislators’ room for transpositional manoeuvre.

Yesterday over lunch¹ i attended an online seminar on “Single Market 2.0: the European Union as a Platform” by Andrea Renda (CEPS). Despite the rather silly “2.0” moniker in the title, his presentation turned out to be a veritable tour de force through the digital policy space in general and the digital policy issues currently on the agenda of the EU Commission. There is a lot to unpack and to learn here but for now i leave you with Renda’s sketch of his proposed architecture for the the Single Market 2.0:

For this to make any sense to you, you will either have to watch the recording of the seminar or read the underlying research paper.

To document the work on our new reserach project on interoperability Alek has started a bi-weekly newsletter: The interoperability. In the first edition he reflects on policymaking as a form of worldbuilding. I particularly like this passage:

Policymaking, the crafting of reality through strategic documents, is in this context a deeply humanistic thing. If you try to build worlds with policies, there’s a high chance you believe that social norms and contracts, and democratic oversight, can trump technologies.

Go read the rest and subscribe here.

I quite like how this map by Arturo Muente-Kunigami displays the tensions between the different conceptual approaches to data displays the inherent tensions between them:

The map is taken from an essay advocating for a more holistic approach to data. Here is the bit where he describes the different approaches/camps:

Take, for example, the Transparency and Access to Information camp, demanding that virtually all information should be made available to the public if requested (with few exceptions like national security and personal information). Close to their tent is the Open Data camp, requiring a proactive publication of all data (with similar exceptions) in machine readable formats and with an open license to both improve transparency and accountability while promoting innovation. We also have the Data Protection cluster, primarily concerned about the potential violation of rights and discrimination that could arise from unregulated collection and use of personal data. Next, a growing group of Technology Enthusiasts, mostly from the private sector, promotes the use of all data available to fuel new technologies (notably, artificial intelligence, big data analytics, and automation) under the promise of a more effective and efficient delivery of goods and services. This group includes a subset of more aggressive vendors buying unstructured data to target consumers with personalized proposals. Digital Government Authorities are promoting the concept of “data sharing” (interoperability) so that citizens don’t have to present information to one public agency if another agency already has it. And close by, Cybersecurity folks are prompting everyone (governments and private companies) to put in place security measures, highlighting the potential threats that arise from the mere existence of these huge pools of data and the harm that criminals could inflict to the economy should they get their hands on them. There sure are other groups that have not been listed.

The main problem with this description is the use of the term “tech enthousiasts” instead of the much more appropriate “surveillance capitalists” for describing those who “promote the use of all data available to fuel new technologies”.

This reminded me of how we have been working at COMMUNIA for the past few years:e

Split the information you’re consuming amongst your team, whether that team is your colleagues, family, friends, or your professional fellow travelers. Read different sources, sorted by local and global, social and technical, known reliable correspondents or experts, and so on. Sort the light from the heavy, the narrative from the data, or whatever way works best to keep you from carrying the burden of all of that cognitive load alone.

Individuals build specialized knowledge, develop good critical radar for sources, and track rises and falls in information trends quite well. A group with diverse interests, tolerances, knowledge of industry dialects and players, strategically spread across time zones, can work together to keep their eyes on the landscape, and save any one person from burning out on breadth, depth, and span of material. This approach can also help mitigate overly negative or overly positive views of one person.

It is a passage from a longer essay titled “Managing Abyss Gaze In a Time of Difficult Futures” written by Scott Smith that is pretty insightful (if you can tollerate the pretty thick layer of post-consultancy lingo). (via BruceS)

For a while techdirt has been doing an admirable job at documenting the failures of the copyright filters. This week has been a particular good week with two new cases that demonstrate the shortcomings of filtering technology (or rather automated systems build on top of such technology):

On Tuesday they reported about NBC universal and National Geographic claiming to own NASA footage of the recent launch of the SpaceX Crew Dragon and as a result taking down YouTube clips and streams of anyone else using this footage. The issue here are incorrect rights claims that even in obvious cases like this one are not caught by YouTube. As long as the entities making such claims have to face no consequences for such behaviour this will continue to be a side effect of the deployment of any automated filtering system.

Now preventing others from showing images of an overhyped commercial spectacle is relatively harmless from a freedom of speech perspective. Today techdirt reports about another case that is almost the textbook example of copyright filters blocking political speech: YouTubes Content ID has blocked reporting from the ongoing anti-racism protests in the US because it contained music by 2Pac and Marvin Gaye playing in the background. This well documented failure mode of copyright filters should rule out their use under almost all circumstances.

Thanks to techdirt for continuing to remind us that the issues we have been highlighting in the discussions around the implementation of Article 17 of the Copyright Directive are not merely theoretical but very real problems that even the self styled leaders in content identification have failed to solve despite massive investments into their technology.

Today the Commission published its long awaited public consultation on the Digital Services Act [announcement , pdf consultation document]. While such consultations are certainly intersting and answering to them is important, we should also not over estimate the importance such consultations play in shaping legislation. Historically the EU’s track record with listening to consultation responses is rather bad. In the case of the copyright directive, where the 2016 proposal was preceded by a broad public consultation in 2014, the responses to the consultation where largely ignored and the issues that would turn into the most controversial parts of the legislative proposal where not even part of the consultation.

Instead, while civil society and academia spend their ressources on responding to the consultation (and in other public fora such as hearings), the entertainment industry was busy pushing its own agenda and narratives via closed door meetings with high ranking commission officials.

To shape the outcome of the discussion on the Digital Services Act, we will need to make sure that we are not too focussed on public feedback. Instead we will focus on building narratives and reach out to the decision makers on the political level.

Speaking about vertical video: This NY Times feature about street skating in New York City during the lockdown really resonated with me:

The quotes from these skaters should be required reading for anyone trying to understand the concept of public space.

More evidence that Instagram TV is encroaching on YouTube as an online outlet for creators. The Verge is reporting that Instagram will share revenue with creators for the first time through ads in IGTV. The article does not really go into my previous question of how Instagram will be dealing with the use of copyright protected third party content in user uploads, but it seems that for those videos and creators that will be allowed to monetize their uploads they are taking a better-safe-than-sorry approach. They will manually screen all uploads before allowing monetisation:

Osofsky says Instagram is human-reviewing every IGTV video before it’s accepted for monetization during this first phase. Eventually, the hope is to deploy a combination of human and software review. He says Instagram will rely on Facebook’s moderation work in terms of both the “technology and the people that do the reviews.”

It will be interesting to see if Instagram manages to build a less controversial system here that can scale beyond he limits of human review, or if they will be happy to keep this limited to a set, of hand-picked, advertiser-friendly, copyright law-abiding creators.

I also find it fascinating that this move by Instagram is potentially another stepping stone in the rise of portrait-orientation video replacing landscape-orientation video. As the Verge notes the …

…introduction [of monetizable ads] sets Instagram to compete directly with YouTube, especially if creators start prioritizing their vertical video shoots over landscape and brands spend their money on Instagram ads rather than YouTube.

Given that Instagram has already surpassed YouTube in total advertising revenue¹, creators have very good incentives to do so.

So for some reason a request from someone in an unnamed EU canditate country to review amendments to their copyright law implementing the 2014 EU Orphan Works Directive landed in my inbox. It is a genuine request from an undoubtably well meaning person, but after calculating the likely impact of the OW directive, i just cant bring myself to do this:

Since its entry into force in 2014 the cultural heritage institutions in the 28 (former¹) member states of the EU have identified a grand total of 12.847 orphan works and registered them in the EUIPO database. That is 0,000025 Orphan works per inhabitant of the EU28. If we take population as a proxy for the capacity to identify orphan works then this would suggest that the 4 western Balkan candidate countries will identify a mere 305 oprhan works in the five and a half years after having implemented the directive (ranging from 16 works in Montenegro to 177 in Serbia).

It is pretty cruel and shameful that we make accession countries implement failed legislation that will do them (or anyone else) no good whatsoever as a condition for joining the EU. It is one thing to keep failed legislation on the books because no-one has the guts to publicly admit to its failure², but it is another thing to subject ressource-strapped countries to implement this legislation just because it exists. Someone should take note of this and get rid of the directive before it wastes even more ressources without doing any good.

Last week the US copyright office released it’s long awaited (in copyright circles at least) Section 512 Study¹. Section 512 of the US copyright code codifies the notice and takedown rules for Online Service Providers (OSPs) that were established by the Digital Millennium Copyright Act (DMCA) in 1998. In the area of copyright this is essentially the equivalent of the Article 14 of the e-Commerce directive that has been at the center of much of the controversy surrounding Article 17 of the Copyright in the Digital Single Market directive.

The Section 512 Study is a first step towards a possible legislative reform of the US rules that the entertainment industry has been lobbying for since the DCMA. At first glance the study stops short of fully endorsing the entertainment industry’s wishlist: It does not include Article 17 style requirements for OCPs to make best efforts to prevent the availability of works (i.e an indirect obligation for upload filtering), or to implement notice and staydown systems.

On the other hand a lot of the findings seem to undermine the legal position of online platforms (and their users²) and strengthen the position of rightholders. On the Kluwer copyright blog Pam Samuelson provides a first assessment which contains this passage:

The biggest disappointment for me in the Section 512 Study was its cavalier and largely dismissive attitude toward fair use. The Study discounted evidence of wrongful takedown notices offered by OSPs and civil society groups. It criticized as wrongly decided the Ninth Circuit’s decision in Lenz v. Universal Music Group Corp. [] (9th Cir. 2016). Universal sent a takedown notice to YouTube regarding Lenz’s short video of her baby dancing with some Prince music in the background, which Lenz thought was fair use. The court in Lenz held that a copyright owner could not claim that it had a good faith belief of infringement before sending a takedown request if it did not consider whether a use such as Lenz’s might be fair.

The Study notes that some OSPs have decided not to remove content alleged to infringe when the OSPs were convinced the challenged uses were fair. The Study chided them for doing so: “OSPs do not appear to be fully honoring the requirement in § 512(c)(1)(c) that upon receiving a takedown notice that is compliant with § 512(c)(3), they ‘respond[] expeditiously to remove or disable access to’ the material.” Under the Office’s interpretation of § 512, in other words, OSPs must remove or block access to content about which a takedown notice has been received regardless of whether the use is fair.

This approach would render fair use a meaningless protection for freedom of expression on online platforms. The attempt by rightholders and copyright maximalists to frame exceptions and limitations to copyright as privileges that can only be claimed ex-post (be it in the form of the DCMA counter notice process or the new complaint and redress mechanism established in Article 17 of the DSM Directive) has the potential to do serious damage. While it remains to be seen if this new attack on fair use as a ex-ante users’ right protection will be successful, it is worth noting the the EU system with its specific exceptions may ultimately provide stronger legal safeguards here³.