Insulting messages from strangers, including spelling mistakes, will soon end up in the trash.
The social network Instagram is gradually setting up from this Wednesday a new filtering tool for content and offensive words sent in invitation requests by message.
Algorithms will automatically filter these missives thanks to a predefined list of expressions, words but also emojis diverted from their proper use.
"We have chosen to apply this filter on connection requests between accounts because these are where offensive messages arrive according to the feedback from our users," says Clotilde Briend, the head of public affairs at Instagram France.
The same method of filtering hate comments that has been in place since 2017 will come into action.
And that promises a hell of a cleanup.
READ ALSO>
Three questions about ten years of Instagram
Concretely, the message of a user who insults another "in DM" without following it will be redirected to a parallel inbox where the filtered words will be blurred.
The person concerned is then free to consult or delete them.
"This gives more control to the user while preserving the confidentiality of these private exchanges to which we do not have access and which we do not proactively remove upstream," says the manager of Instagram France.
An option requested by French influencers
This option had been demanded for a long time by public figures, in particular French, who were regularly the target of stormy interactions in their messaging. At the same time as the Germans or New Zealanders, French users will be among the first in the world to experience this new tool. Its deployment could take place "within 2 to 3 weeks because the list of words to be filtered drawn up with the association Génération Numérique is not complete". At the global level, the filtering will be effective during the month of May.
Also as part of the fight against online harassment, the social network owned by Facebook announces another tool that will automatically block the accounts - including newly created ones - of a user reported for abuse or hateful content. Any new accounts of a user suspended for breaking network rules will also be deleted as soon as they are detected.
