The advent of Big Data made this planet a better place to live in but there are also downsides.
Recently (currently…) I have to suffer my second suspension of my Facebook account for an alleged breach of their Community standards. As a lawyer I know that these Community Standards are not as their name suggest: norms made by the community but simply contractual terms and Facebook has every discretional power to decide what is against these standards. No problem, when I signed up I also signed up for this.
However, this temporary block is a good example to observe every possible downside of Facebook’s BigData-driven approach to keep the Facebook a safe place.
The post I was blocked for contained the Hungarian term “büdös zsidó” (stinkin’ Jew). The context was that The Worldwide Jewish Plot, a common subliminal hallucination in the Hungarian political community is always a refferred to as a flock of filthy, stinky, evil bastards whenever they give money to causes like liberty in equality. In striking contrast, they are a cool gang of wise guys when they build a fence on the Israeli border like the Hungarian government does our Southern border. The Fence is the symbol of the Hungarian politics. Even the opposition admires it and anyone speaking against is bound to face destruction and oblivion. Except a few people. Including me.
Now, what has this do with Facebook’s Community Standards?
First, I am Hungarian. Like many Central and Eastern European languages, Hungarian builds a lot on secondary, terciary, quaternary …zillionnary meanings. If you run into an expression like “filthy Jew” you are either on nazi extremist page or someone might just use the word to illustrate something. For a native speaker the distinction is easy but for a computer it is close to impossible.
My language is spoken by around 10 million people. Maybe half of them have a Facebook account. Compared to the billions who use Facebook in English this is nothing. This might be the reason why Facebook never cares to build a proper analytical software and/or use actual living people to check what the system is doing.
Why is this a problem?
Hungarian media was taken over by the government in the recent years. The rules and the scope of political debate are set by the ruling Fidesz party.
Anyone who wants to operate outside these lines and also find some audience must go on Facebook and set up a page. This is great news, many Facebook pages gained big audiences.
This is why we showed up on the radar screens.
Couple of days ago András Jámbor, editor of Kettős Mérce (double standard) a leftist-liberal FB page was banned for a week and in the same time it is my second suspension for the same reasons.
And here we get to the downsides of Big Data. If a system is not overly sophisticated (like the Hungarian language analytical system of Facebook) it is quite easy to spot where it jumps to an automatic suspension.
One only has to find the nipples and cameltoes in Hungarian and then report them enough times to set off an automatic block. Someone just did it. And this someone is the same one who gradually restricted free Hungarian media to a couple of Budapest-based webpages.
Now, because of some lousy coding work they have a weapon against the rest. They have a weapon against the biggest achievement of the Internet: freedom of speech.
Ok, one might use substitutes for certain words. We, Hungarians are masters of code speech, centuries of suppression taught us how to do it.
The thing is that we want to forget it. This way, we never will.
As I am unable to share this on Facebook, I ask our readers and fellow editors to do it.