This article tells us about Germany's plan to fine social media sites over hate speech:

Germany plans a new law calling for social networks like Facebook (FB.O) to remove slanderous or threatening online postings quickly or face fines of up to 50 million euros ($53 mln).

"This (draft law) sets out binding standards for the way operators of social networks deal with complaints and obliges them to delete criminal content," Justice Minister Heiko Maas said in a statement announcing the planned legislation on Tuesday.

We are also informed about the rate of illegal posts removal:

A survey by the justice ministry's youth protection agency, released on Tuesday, found that YouTube was able to remove around 90 percent of illegal postings within a week, while Facebook deleted or blocked just 39 percent of content deemed criminal under the law and Twitter only 1 percent.

This study highlights the ambiguity surrounding hate speech (might also apply to other illegal type of speeches):

For the purpose of training a hate speech detection system, the reliability of the annotations is crucial, but there is no universally agreed-upon definition.

Our results indicate that showing users a definition caused them to partially align their own opinion with the definition but did not improve reliability, which was very low overall. We conclude that the presence of hate speech should perhaps not be considered a binary yes-or-no decision, and raters need more detailed instructions for the annotation.

A classical example shows how much a text can be scrambled and still be readable (and does not even use digits and other characters instead of letters), making automated recognition quite difficult. Of course, natural language processors can be improved to cope with such texts, but users have a large freedom in distorting the text.

Question: considering above information, can Germany enforce such a law?

  • Why not? Your second link suggests only that they have plenty of work ahead to make it quick and automatic, but I would not expect a German judge to be someone an international player can safely ignore.
    – user9389
    Commented Mar 14, 2017 at 20:24
  • 1
    This appears to be a technology question more than a political one.
    – user1530
    Commented Mar 14, 2017 at 20:52
  • 2
    Keep in mind that it's currently just an idea of a minister, not a fully written out law. And it's an idea brought up in an election year. So it's not clear how serious this plan should be taken.
    – Philipp
    Commented Mar 14, 2017 at 21:03
  • 1
    This is about the law not the political processes. I think if you limit it to if there is anything in the German or EU Constitution that would allow it that would be fine. But the scope is currently off topic Commented Mar 14, 2017 at 21:26
  • 1
    @Alexei: I think software developers often have a real problem with "lack of clarity", but judges, especially German judges, don't. Just because you think it's hard or impossible to decide what's hate speech doesn't mean a judge cannot decide. BTW. The US "beyond reasonable doubt" refers to whether you did the act what you were accused of doing or not. It doesn't refer to your act being illegal in the way it is suggested or not.
    – gnasher729
    Commented Mar 15, 2017 at 21:54

1 Answer 1


It might be relevant to read the summary of that proposal on the website of the German Ministry of Justice. I can only find the original German press release and the actual text of the proposed law (also in German) on the website of a lawyer. A short summary of the summary:

  • The only content which needs to be removed is content which is violating existing criminal laws, like slander, insults, incitement, threats, etc.. So it doesn't make any speech illegal which wasn't already illegal before. What's new is that any "social network" (which the proposal defines as any website which primarily features user-generated content and has more than 2 million registered users in Germany) will be punished if they don't remove that content, and not just the user who posted it.
  • "Social networks" only need to delete things after a user reported them. So there is no need for automatic content filtering.
  • "Social networks" must have some convenient way for users to report illegal content. So crowd-sourcing the search for illegal content is not just an option, it's mandatory.

A problem which I see with this proposal is that websites now have the responsibility to decide by themselves what's illegal content and what isn't. When the ministry decides that they didn't delete something they should have, the case goes to a court. And when the court says it should have been deleted, the website is fined. So a misjudgment by a moderator could cost a hefty fine (up to 5 million €, in fact).

It's also a bit questionable how they intend to enforce this law. It is obviously targeting international internet companies like Facebook. The monitoring report from non-governmental organization Jugendschutz.net which is linked to the press release specifically calls out Facebook and Twitter for having too lenient moderation (Youtube is allright, according to the report). But Facebook doesn't have a representation in Germany. The press release says they want to somehow force international companies to name a representative in Germany to collect fines from, but not how they want to do that. The only options I see would be to either get the International community (or at least the EU) to cooperate and enforce the German demands against these companies, or to threaten these companies with being censored from the German internet. But the last attempt to build an internet censorship infrastructure in Germany was the "Zugangserschwerungsgesetz" in 2009 which was received so negatively by the population that it was never enforced and eventually repealed (and in that case it was about blocking child porn, not blocking websites which millions of Germans use every day).

Keep in mind that this is currently a proposal fresh from a ministry. It was not yet passed through a parliament commission. Law proposals from ministries can often change a lot before the parliament votes on them. Also keep in mind that 2017 is an election year. If the Ministry of Justice changes hands in 6 months, that proposal might disappear without a trace.

Update March 15th 2017: I found the actual text of the law proposal and corrected a few things.

  • 1
    Very nice answer. Another problem I see is "Social networks only need to delete things after a user reported them. So there is no need for automatic content filtering.". If a social network collects reports, it can just tell that no one has flagged the post so far. So a user will need to flag a post and produce the recording that he flagged a post :-). But overall I find this funny: you do not have a representation from which we can take your money, so create a representation so that we can take your money. Commented Mar 14, 2017 at 22:24

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .