Article 13

Created: 2018/07/01 21:43:24+0000

Article 13 of a directive on copyright introduces automated scanning of user uploaded content to identify copyright violations. This includes text, images, audio and video. People make accusations of censorship. The directive calls on the relevant parties to define best practices and explicitly mentions "appropriate and proportionate content recognition technologies" as an example. Content recognition technologies is where automated scanning comes in. This video explains some instances of copyright infringement that the article might address. Udemy have has been a platform that has been exploited by copyright pirates for years. It's important to remember article would not just apply to sites like Udemy that profit from publishing content, it also affect other sites like Facebook.

There are problems with the use of automated scanning.

  • Copyright places limits on how, what and why things can be duplicated but does not impose a total ban on doing so. Automated solutions don't make determinations based on whether the duplication is permitted by copyright law. This leads to a total ban in practice. Preventing people from engaging in activity that is explicitly legal isn't desirable consequence of a law.
  • It's not clear to me how large a problem this is attempting to fix. Many sites such as YouTube already provide automated scanning. The article fails to address peer to peer systems, where a description of the copyrighted materials is uploaded to central sites. How much would copyright holders really benefit from preventing people from sharing photos on Facebook?
  • The article encourages a technical solution to a social problem. It doesn't consider the technical limitations of such solutions or address implementation details. Do sites need to respect a common registry of copyrighted works? Do they maintain their own registries? Can they charge copyright holders a fee to cover the cost of processing uploaded content? How are people prevented from claiming works they don't own the copyright to like the complete works of Shakespeare? All this is going to be left to the relevant parties to figure out.
  • The limitations of automated solutions may cause sites to implement their own rules, specified in their terms of service, rather than apply copyright law. This could place stringent limitations on user behaviour.
  • Mass automated surveillance seems like an excessive approach to copyright infringement. Remember automated scanning would look at everything uploaded and evaluate it. You're taking a machines word that someone is violating copyright law. If content is publicised before the automated scanning is completed the uploaders could still be exposed to further legal action from the rights holders.

For a deeper break down of article 13 take a look at this document by EDRI.

Much of the consequences of the article are likely to depend on how "appropriate and proportionate" is interpreted by companies, courts and the member states of the EU. The article appears to drafted to cause long running debates rather than providing clarity.