Social media moderation: RTBF’s daily challenge

Banning, on the other hand, is reserved for the most serious cases: repeated insults directed at journalists or presenters, death threats – which are escalated to RTBF’s legal department – or automated spammers. There is also the mute function: the user can continue to comment, but nobody sees their messages. “People aren’t banned; they can continue to comment, but nobody sees their comments any more,” explains the coordinator. It’s a solution that prevents the situation from escalating whilst neutralising problematic content.

Emojis to get round the filters

Moderation isn’t limited to spotting obvious insults. Some users have developed a veritable code language to slip through the cracks of the automatic filters. The example of “arbres”  (« trees » in english) illustrates this well: on a post commemorating the victims of the Bataclan, the word “arbres” — a substitute for a discriminatory term — must be identified and hidden. Some even go so far as to use a simple tree emoji.

“Those who comment know full well that if they write the (discriminatory) word in question, it will be hidden. So they come up with ways to get round the moderators,” explains Aline De Volder. “Moderators have to adapt each time because users will find new ways to get round the moderation system.” Lists of keywords that trigger alerts are regularly updated to keep pace with these constant changes.

“The principles of public service clash with the platforms’ ambition to generate as much money as possible through advertising” – Damien Van Achter, media consultant and head of innovation at IHECS

It’s not always easy to decide: citizens share their views

This is one of the highlights of this “Table Média” session: participants are presented with real-life moderation cases and opinions differ.

A harsh comment directed at a mother in a vulnerable situation divides the group. Should the comment be left visible to fuel the debate? Or should it be hidden immediately? Some members of the public lean towards freedom of expression, whilst others feel the comment crosses a line. The discussion reveals just how blurred the line is and how differently people can react to the same text.

“Moderation, at times, is subjective,” admits Aline De Volder. “Different people may not approach moderation in quite the same way.” To minimise these discrepancies, RTBF relies on guidelines tailored to each account. The rules are not the same on a news account as they are on Views or Tarmac. Felice Gasperoni, from RTBF’s commissioning department, emphasises the importance of this collegial approach: “Among ourselves, we have a discussion thread where we sometimes share our doubts, saying, ‘Hold on, what should we do?’”

A peaceful discussion environment

Faced with the excesses of the major platforms, RTBF is working to develop interactivity on its own platforms – its website and, eventually, RTBF Auvio. The idea is to offer a more peaceful discussion environment, far removed from the dynamics of social media. Initial tests are encouraging. “When users interact on our platform, they naturally adopt a different mindset,” says Aline De Volder. This observation is welcomed by several participants around the table, who are weary of the excesses they witness daily on social media.

Damien Van Achter, a media consultant and head of innovation at IHECS (Institut des Hautes Études des Communications Sociales), urges us to look at the root of the problem. “The principles of public service clash with the platforms’ ambition to generate as much money as possible through advertising,” he analyses. On X in particular, documented studies show that the algorithm deliberately favours divisive content. “X’s algorithm highlights content that we know people will get worked up about,” he summarises. For the participants in “Table Média”, this observation reinforces their conviction that preserving digital spaces underpinned by public service values is not a luxury but a necessity!