Articles 11 and 13 in the New EU Copyright Directive Enable More Surveillance

Privacy activists were caught off guard last week when the EU passed its controversial copyright legislation. The new law has far reaching consequences that technical experts and internet companies alike have strongly advised would damage EU businesses and change fundamentally how the Internet works.

This article is intended to show you WHY these two directives (articles 11 and 13, known as the link-tax and upload filter) are highly problematic from a technical point of view, and how these laws will further weaken privacy on the Internet.

Article 11 – The “Link Tax”

The intention of the law (AKA how the EU thinks this will work):

The EU is working to protect media companies from piracy of their content. Sites will have to pay a fee to the creator in order to link to their content, creating a new revenue system for waning traditional news media while giving them an opportunity to get compensated for valuable content.

What likely happened here is a bunch of people who do not understand how the Internet fundamentally works have looked at statistics on how many times Google, Yahoo, Reddit, Twitter, etc have linked their content and multiplied that by some arbitrary number for their “link-tax” and decided that they can now cash in because of their new legislation.

The technical barriers and the likely reality:

The above explanation completely misunderstands how the Internet actually works.

The search engines and social media sites will downrank EU content to make it appear infrequently, or remove the content entirely from their sites. (This already happened when Spain tried this idea before.) This will cause an enormous drop in traffic to their content (which will dramatically decrease ad revenue) that will not be offset by the revenue from the Link-Tax. In essence, the EU will be damaging its own media companies and empowering outside media companies who will not carry the same technical and financial burdens.

This also creates a problem in that “copyright trolls” can claim content that doesn’t belong to them and create problems for creators, as well as creating a huge backlog of disputes to be reviewed manually at tremendous costs for the platforms that are serving EU content. Only the largest companies will be able to carry this regulatory burden.

The new law does carry an exemption for “small and micro businesses” but makes no effort to clearly define what a small business actually is nor what types of businesses are exempt.

Article 13 – The “Upload Filter”

The intention of the law (AKA how the EU thinks this will work):

This law requires that a system be in place to quickly check all uploaded content to a site or service be checked for “copyright infringement”, and that any infringement found needs to be barred from being posted to the site. It also specifically calls for a fast appeals process that allows false positives to be corrected. The idea is to create an environment where it is hard to pirate content on any major sites or services in the EU.

The technical barriers and the likely reality:

It is hard to find where to begin. This law is so ill conceived that nothing about it is redeeming and it shows no understanding of technology at all. Major sites like YouTube, Twitch, SoundCloud, and Vimeo all get YEARS of content uploaded every day. Lawmakers say that a filter is “not a requirement” but there is no other way to comply with this law which brings us to the first  major problem…

Even the big tech companies cannot create good filters. They’ve tried.

YouTube is the biggest example of attempting to use an automated filter to control copyrighted content. It’s algorithms can catch audio and video of 1:1 pirated content, and some types of other content. These systems have big technical problems that will not be solved any time soon. They remove content with no regard for fair-use, they have high false positive rates for unrelated content, they often remove content with incidental content in the background (a copyrighted song playing in the background, etc), and these systems are heavily abused today by copyright trolls.

And this is the BEST that the biggest companies can offer.

Like article 11, article 13 seeks a fast appeals process. This will place an unbelievable burden on sites that host a lot of content for continuous manual review of every dispute that surfaces. The costs involved here will damage companies like YouTube that are already losing money or breaking even.

Further, even these best efforts at filters are easily defeated. They often work by sampling areas of a video looking for matches or matching audio signatures by looking at pitch, beats per minute, and other basic audio characteristics. Defeating these filters often involves just a 3D tilt of the video, slightly changing the colors, or adding subtle grid-lines or dots to the content. Audio filters are frequently defeated by speeding up or slowing down the audio or shifting the pitch. The technology to defeat these kinds of efforts requires computing power far beyond what is reasonable and again leads to a cost problem for businesses that even the well-known giants will have problems developing and operating while remaining profitable.

This leads us down the rabbit hole of what smaller sites will be able to do. There’s no question that startups and fledgling businesses will struggle to develop their own tech for this. So what is left? The Internet giants will gladly lease you a solution at a tremendous cost burden, or they will allow you to use the tech for free, in exchange for seeing all of the content posted by everyone worldwide. Make no mistake, this regulation is perfect for the Amazons, the Googles, the Facebooks and the Microsofts of the world. They will develop filters that spy on every creator inside of the EU, and they are forced to give up their data, or the sites have to pay a heavy financial burden.

Ethics Concerns: Articles 11 and 13 Give No Regard to Censorship Either

While the new copyright directive does much to place burden on companies to police content and provide an appeals process, it does little to punish people or organizations that make false copyright claims. This will lead to censorship of people or groups and shield bad actors from criticism, because the groups can simply claim copyright over quotes, images, logos, or even their own faces to silence critics. One can easily imagine scenarios such as playing a song loudly at a major protest to justify the removal of all audio of said protest online, or calling for removal of all video that contains a particular building or landmark.

Applicability: All 28 EU Countries Develop Their Own Rules

Even with all of the above problems and concerns, this is further worsened by what will be an inconsistent regulatory environment. If Germany writes their version of the law with exceptions for Parody and Political Action, but Poland does not, and Spain only allows two words from an article’s title to match before a Link-Tax must be paid, you have a gnarled regulatory environment that no one can navigate on top of your already unwieldy and technically unworkable copyright legislation.

Purpose: It Will Not Stop Piracy

The core issue when looking at the technical and ethical problems is that this isn’t even a trade-off. This will not prevent piracy in any measurable way. It is all easily defeated and piracy sites will not follow any of these rules anyway.

A Slippery Slope to More Censorship

This is a first step in laying down infrastructure that can silence opposition, remove criticism, and filter the internet not just for copyright, but all types of dissent. The first steps in this direction lead down a long path of increasingly terrible consequences. Opening Pandora’s censorship box will lead to more problems in the near future as nations decide what new things to block based on how inconvenient they are for the hegemony.

It is absolutely crucial that this legislation must never come into force. It will not serve any of its intended purposes and will place a tremendous cost burden on companies that will stifle innovation while giving no real benefits to content creators. This idea has failed before and will continue to fail with this effort only on a larger scale and more spectacularly.

Stopping Censorship and Preventing Damage to the Internet:

To help, find out which MEP’s in your country voted for this legislation, (link requires signup but no verification) and find out why. Call out the bad actors that are trying to protect dying industries with bad laws that hurt global commerce. Do not let lobbyists decide the future of the Internet.