CopyrightX: Article 13

Panel Two: Online Service Providers, Automated Anti-piracy Systems, and Article 13 [like CDA 230, it may never lose the initial number even if it no longer fits the statutory numbering]

Moderator: Jennifer Esch

Giancarlo Frosio: Copyright and Article 13

Summary of current provisions.  Automated technology is the intended effect. Hard to see how that can be done without general monitoring of all contents of a platform for unlimited times.  Scarlett court made clear that fundamental rights are implicated—freedom of information and expression—and automated technologies are not proportional, especially given that they might jeopardize a business.  Limitations and exceptions also need to be accommodated somehow.

Ginka Hristova: The Effects of EU Article 11 and Article 13

Content ID is the prime example that people give of how it should be done but Content ID works badly—it’s not a good model.  [And the same content owners complain bitterly about Content ID not being good enough!]  Burden of proof is reversed from the Electronic Commerce Directive.  Now platforms will have to prove that they’re making enough efforts to avoid direct liability.  Collisions w/GDRP as well.  Especially when determining whether exceptions and limitations apply, much information will need to be collected.

Mohammed Iriqat: Copyright Piracy in the Arab World

Extensive unlicensed downloading of music, movies, software. 60% of software running in the Arab world is unlicensed.  Millions of DVDs.  Gov’t institutions and universities participate.  Microsoft Office $5 an unlicensed CD, compared to $70 authorized price.  $2 billion in calculated economic loss.  Possible solutions: collective management organizations, but there are just 5 Arab CMOs and 22 Arab countries. Raise awareness. 95% of Arab citizens don’t see a problem with illegal downloads.  Internet laws like DMCA, which are lacking in the Arab world.  Customs authorities: no control over entry of pirate CDs.  World Cup 2018: Four countries blockaded Qatar, which had the rights to broadcast; they made sites to distribute the World Cup w/o authority of the Qatari rightsowner. Intergovernmental conflict.  Qatar went to WIPO for arbitration, but WIPO hasn’t made a final decision.

Jordan Gimbel: Twitch and the Likely Effects of Article 13

Wide spectrum of online service providers and treatments of unauthorized use/debates w/ different rightsholders: music, film, gaming.  Twitch isn’t YouTube or Twitter; has some features but also Patreon-like.  You can live broadcast your gaming.

Streaming could infringe, but game publishers/developers often license streaming—an attractive complement to their titles.  Many license monetization by individual streamers. But we’re not immune from unauthorized works appearing.  We do see notices for (1) livestreaming sports; (2) violations of embargo conditions on new games; (3) user disputes over emotes (users who create designs for the overlay).  Twitch as complement: NBA brought its G-League (farm team) cames to Twitch, allowing others to costream it (add it to their own channel with their own overlays).  NFL selected a few costreamers who could stream it to their own audience w/their own commentary.

We process DMCA notices in minutes; suspend live broadcast for 24 hours. For VOD works we mute the songs—if you create an archive of a live broadcast, we scan & create audio fingerprints w/Audible Magic and mute portions for matches.  Some rightsholders have priority access—soccer leagues, game companies—to act w/o needing to send notice.

How does this change under Art. 17/13?  Obligation to license, notice and staydown: filters are the only option.  Reference files aren’t available for livestreams so that’s a challenge. This content is yet to be created.  And the overlays are different digital assets. We can scan the audio, but scanning images on overlays from multiple sources are a challenge. Scratching our heads.  Algorithmic tools are going to be needed; taking $ from other developments. Open Qs.  Mistakes in livestreams lead to suspension of channel, which impacts people’s livelihoods.  Filter’s false positives are therefore concerning. Crossborder implementation challenges: different rightsholders across Europe.  GEMA (German PRO) announced that it thought it should be given a larger negotiating power on behalf of images as well as music. That’s great, but only for Germany.

Q: what could gov’ts do/what are the missed opportunities?

Hristova: implementation legislation.

Q re real people depicted in games.

Gimbel: US-centric answer: right of publicity is what’s debated.

Q: where do you get the filters from? Google?

Gimble: question of market power.  Content ID has been around for a while and they’re set on approach. Compliance w/best efforts: we’re trying to figure out whether that means we can engage in licensing from a vendor, but may have to develop a filter that maps to your service using your reference materials. Open Q: does the law require us to build our own tech to do that because we have a different way in which we serve content, w/overlays that are different from other services. [One of my biggest questions about all this is how licensing is going to work for visual works—is there just going to be a wealth transfer from platforms to the small percentage of visual artists who are represented by PROs and to the PROs themselves based on the overall “value” of visual works?  Whereas music really does have a high percentage of covered material deliberately part of a PRO, visual works don’t and won’t.]

Terry Fisher: some suggestions about implementation: make platforms responsible for costs of wrongful blocking of noninfringing materials in order to incentivize balancing and use of mechanisms such as arbitration or option system to reduce social costs of this regime.