23 August, 2018
You were kind enough in the past to comment for SPUTNIK International, tech issues on the protection of your personal data. Your CODIFIC expertise would be greatly appreciated to comment on the decision by Commissioner Julian King to impose a maximum of one hour to tech companies for taking out of the net, the terrorist messages. Just a few questions…
Can these organizations take the “necessary technical and organizational measures” to reach that goal? Facebook, Twitter and others need to identify the terrorist message and then take it out. Not easy, is it? Is it feasible?
There are quite some censorship studies of Twitter during Turkey coup and Weibo in China that have demonstrated that this should be feasible and scalable*. I guess the recipe is to use natural language processing combined with some heuristics for the metadata (e.g., user profiles, post popularity, location, etc) and human moderators should yield pretty good results. Of course, there will always remain a certain percentage of legitimate messages identified as “terrorist messages” (i.a., false positives) as well as “terrorist messages” identified as legitimate (i.e., false negatives). So yes, this is feasible. However, it will have an impact on budgets and tech giants are most likely not really excited about this.
* “The decline of social media censorship and the rise of self-censorship after the 2016 failed Turkish coup”, R. Tanash, Z. Chen, D. S. Wallach, and M. Marschall).
If feasible for the giants, what about smaller players that don’t have the human or technical resources to scan their domain?
I don’t think small players are in scope. For them it is not really feasible.
German authorities and others seem to be more annoyed by “hate speech” towards them. Fines have already been imposed for that. Should the EU set priorities: first terrorism within an hour, and then only think about fines for ‘hate speech”?
Again, technologically this is feasible, however, where exactly do we draw the line between censorship and hate speech?
Nobody is “for” terrorism, so why on earth would the big techs try to create themselves problems with the authority, Mr. King, by being sloppy on terrorism?! Isn’t Julian King making himself some cheap and easy publicity ahead of Brexit? Is it part of his career plan?
Well, to “clear terrorist messages” will be quite costly and at the end of the day big techs have the mission to make their shareholders happy by producing a better balance sheet than the previous quarter. On top of that, I really doubt that simply cleaning terrorist messages from social media will deliver any tangible results. If someone goes around the neighborhood and writes terrorist messages on every building / house, your focus should be on catching the suspect and not deleting the messages within the hour.