2023-07-24 17:49:26
It is “often with some success” that social networks have devised, in recent months, solutions to “moderate and mitigate the most obviously harmful misuses of their services”. This is one of the conclusions, in the form of encouragement, of the last with of the Audiovisual and Digital Communication Regulatory Authority (Arcom, ex-CSA) on the means implemented by these operators to fight against online hate.
Written based on observations from Arcom and statements from thirteen platforms (Twitter, YouTube, Dailymotion, TikTok, or even Instagram and Facebook, from the Meta group), this report was made public on Monday July 24, one month from the entry into force of the Digital Services Regulations (RSN, or DSA in English). This will be next August 25 for the largest platforms, and in February 2024 for those with less than 45 million European users.
End of August, “the moment of truth”
In presenting this report to the press, the president of Arcom, Roch-Olivier Maistre, insisted on the fact that it seems “in a period of transition” : “We are in the process of moving from national regulations to European regulations. » Until now, in fact, social networks were self-regulating, possibly framed by the first legislative texts in certain Member States. In France, the law confirming respect for republican principles (adopted in the summer of 2021) regulated the fight against online hate. It will disappear in February 2024, in favor of the RSN.
In presenting their report, Arcom spokespersons reported a “certain feverishness” platforms, apparently concerned with complying with the European text already ” several months ” before its application. The sanctions will, in fact, be much more dissuasive than today: if these companies do not respect their obligations in terms of disinformation, online piracy and hate speech, they can be fined up to 6% of their global turnover. As a very last resort, a temporary suspension of their services is possible.
Waiting “the moment of truth” of August 25, in the words of Roch-Olivier Maistre, what are these operators already doing against online hate? The sinews of war remains the moderation of content. If the social networks are not required to monitor the content published on their services (because they are not considered as publishers, but simple “hosts”), they must nevertheless remove those of them which are manifestly illicit as soon as they have been reported to them.
How many moderators?
For this, reporting tools exist, but access to them is too often conditioned “when clicking on a button whose title is not very explicit”, deplores the Arcom. On YouTube, you have to go to a page called “About”: not very intuitive. As for Snapchat, the user wishing to report an account to which he is subscribed because of illicit behavior must first click on the heading “Manage friendship”!
Another bad signal: most platforms continue to refuse to publicly disclose their number of French-speaking moderators. Authorities and associations suspect these “web cleaners” employed by subcontractors to be too few to cope with the mass of content in circulation. The only ones to have authorized Arcom to publish their figures are Dailymotion, Linkedin, Wikipedia and Twitter. The latter, renamed “X” on July 24, claims to have 149 moderators, “including French speakers”.
Arcom also draws attention to TikTok, which has been in the sights of the authorities for months. When a user challenges a moderation decision, the social network ends up reversing the initial choice of the moderator in 40% of cases. “Is this a sign of overmoderation, poorly trained moderators, or a lack of clarity in TikTok rules? » asks Benoît Loutrel, member of the Arcom college.
The report underlines in any case the “good cooperation” platforms with judicial and administrative authorities: in nearly 85% of cases, these companies respond to requests for information from French authorities in order to identify the potential perpetrator of hateful content online.
#Arcoms #mixed #results