Please Write to Your MEPs to Stop EU Copyright Directive from Seriously Harming the Internet

Next week, a crucial vote will be held by the Legal Affairs committee of the European Parliament (JURI). It concerns the proposed copyright directive, which is moving through the EU's legislative process. Unfortunately, there are two extremely dangerous elements in the current text that will harm the Internet in the EU if passed: basic details about them can be found in this post I wrote for Ars Technica. A third element needs a tweak.

As the Pirate Party MEP Julia Reda explains, it currently looks as if the two bad elements will be accepted by JURI. But the vote is close, and EU citizens have an important opportunity to ask their representatives to influence the outcome of that vote. I urge you to do so, and soon.

You can use the free services WriteToThem, or a new site called SaveYourInternet, to send an email to your MEPs in just a few seconds. The latter site offers some text you can use about one of the problematic parts of the copyright directive, Article 13. However, you may wish to urge your representative to fight against the other bad idea, Article 11. Both of these are explained in the text below, which is what I have sent my MEPs.

Please feel free to draw on this if it is helpful, but it will be more effective if you express yourself in your own words. The most important thing is to send something – no matter how short – asking MEPs to help stop the copyright directive from harming the EU's Internet.

As a journalist who has been covering the Internet for 24 years, I am deeply concerned about the proposed copyright directive that is currently working its way through the EU legislative process. I am writing to ask you to alert your colleagues on the JURI committee to the deep problems with two sections in particular: Article 13, and Article 11. Both need to be removed.

Article 13 will require sites with a large number of user uploads either to license everything they make publicly available, or proactively to stop copyright material being posted. The first option is not practical when dealing with a fragmented market where there is no central licensing agency. And even where such an agency exists, it will not cover every possible upload.

The second option requires sites to prevent unauthorised copyright material from being posted. The only way to achieve this is through a general filtering mechanism. Unless every file is checked when it is uploaded, and compared against a database of copyright material, there is simply no way to know whether it infringes. The fact that a recent JURI version of the directive's text says "The implementation of measures by service providers should not consist in a general monitoring obligation" is irrelevant, because there is literally no other way of achieving the stated aim.

The EU's e-commerce law specifically forbids EU countries from imposing "a general obligation on providers... to monitor the information which they transmit or store." But legal issues aside, there are technical problems too. The upload filters required to block copyright material will be, of necessity, automated – the volume of uploads makes this inevitable. But it is impossible to create a system that encapsulates the subtleties of EU copyright law: even courts have problems navigating their way through this extremely complex field.

As a result, upload filters will be imperfect. The future financial risks of allowing copyright material to be posted means that upload filters will always err on the side of caution, and over-block. This will lead to legitimate material being blocked by mistake. It will have a chilling effect on public domain materials, criticism, parody, and popular Internet memes that frequently draw on copyright material. In short, it will greatly impoverish the EU's Internet, and lead to a massive assault on citizens' freedom of expression. Since licensing is impractical, and upload filters cannot work, Article 13 must be dropped completely.

Despite claims to the contrary, this will not harm the copyright industry. Research carried out on behalf of the European Commission at a cost of €370,000 suggests that unauthorized uploads are not a pressing problem: "In general, the results do not show robust statistical evidence of displacement of sales by online copyright infringements."

The other problematic part of the proposed directive is Article 11, which would introduce an ancillary copyright for news publications. As you doubtless know, this has been tried twice, in Germany and Spain, and failed both times to achieve its aim of revivifying newspapers. It's not hard to see why. The snippets that appear in search engines direct more readers to news sites: they are beneficial for publishers. Trying to force Internet companies to pay for the privilege of sending more traffic to news sites makes no sense. It is no wonder that Google refused to do so in Spain, with serious negative consequences for publishers there.

Some publishers argue that sites are using material from their news publications without payment. There are two situations here. If large amounts of text is being taken, those sites can be sued for copyright infringement under existing laws. If only snippets are taken, as is the case for Google, then this is not infringement, since it is simply using those snippets to direct interested readers to the original article. The snippets are not substitutes for the full text, but tasters encouraging further exploration. In neither case is there any need for additional copyright.

However, if Article 11's "snippet tax" is brought in, it will inevitably lead to fewer links being made to news sites. The public will be less well-informed at a time when misinformation is a growing problem, while publishers will lose visitors. The actual monies from the tax are likely to be small. The German experience shows that very little money is collected in practice. To summarise, then, an ancillary copyright is not necessary, and if brought in will be harmful to the public, with only a tiny benefit for publishers. As with Article 13, Article 11, too, needs to be removed.

Finally, a quick word about Article 3. The idea behind this – to allow text and data mining (TDM) of resources – is excellent. This is a crucial area for things like artificial intelligence, and the EU desperately needs legal certainty here. However, as it currently stands, TDM would not be available to most companies unless they pay additional fees. This makes no sense at a time when the EU is rightly trying to encourage digital startups in the region. TDM will be vital for many services and products, and if companies cannot be assured that they will be able to use this approach when they grow, but will be penalised for being successful, then they will simply set up elsewhere. That is hardly a win for the EU.

The basic rule for TDM is simple: the right to read a text is also the right to mine a text. This means Article 3 needs to be amended to allow any companies, of any size or age, to carry out TDM on texts to which they have legal access.

I apologise for the length of this email, but the topics are complex and important. However, the actions required are very simple: Articles 13 and 11 must be dropped, and Article 3 must be changed. If these amendments are not passed, the effect on the Internet in the EU will be very serious, both in terms of harming the rights of EU citizens, and of discouraging innovation by startups in this region. I therefore ask you to urge your colleagues to make the changes I have suggested.

Thank you for your help in this vital matter.

Seduced by Virtual Elizas

ELIZA was a very simple AI program written forty years ago that nonetheless convinced many people who interacted with it that it was indeed a real person. If we can be so easily convinced by text, what hope do we have against Virtual Elizas in Second Life?

How Cool is Coull.tv?

Today's Web user depends on search - well, I do, at least. But search/Google is really only doing words. And, as any fule kno, words is easy. Now video, that's quite a different matter.

So the appearance of any site claiming to make videos searchable is at least worth a look, so to speak. Coull.tv is one such:

coull.tv enables you to activate objects within a video - making people, objects and other items clickable. Anybody can then add tags and comments describing the video and or the objects in that video. This enables the video or parts of it to be easily found; the more popular tags and comments become, the more often coull.tv will suggest them in related searches. coull.tv will use the power of the community to help categorize and tag every element of a video. coull.tv has no pre-roll or post-roll advertising in the way of the viewing experience.

There are two important elements here. First, the fact that elements within a video can be demarcated and made clickable. But that on its own doesn't make a search engine: it just turns video flow into discrete elements. The second part of the coull.tv equation is to get users to do the difficult bit: indexing all those elements. In fact, this is probably the only way video indexing is going to be done for a long while. Automatic recognition through some kind of AI is just too difficult currently.

God knows, the last thing we need is another video sharing site; happily, coull.tv seems to be searching for something more. (Via John Battelle's Searchblog.)