By akademiotoelektronik, 23/11/2022

Bleep, the Intel anti-harassment filter which allows to pass "a little" of racism

The tool is called Bleep and was presented by Intel on April 8, 2021, as part of the Game Developers Conference.He may make somewhat chatting or, at least, lift a few pairs of questioning eyebrows.

Intended to curb the gangrenne that constitutes notoriously the Hate Speech, racist or misogynist words or coarseness in online conversations of gamers, Bleep is a small live censorship software.

À lire aussiLe «Gamergate» a montré qu'internet était une arme de destruction massive

Installed on the users' computer, it makes its artificial intelligence molded at high speed, designed with the specialist Spirit AI, in order to replace the insulting remarks with beeps.

Motherboard or The Verge report that the video presented by the American firm shows some of the settings that this tool offers: we can ask him to draw up an instant audio shield against validist, racist, sexist, sexual words, against the rough namesbird, the N Word, the LGBTQPHOBES words or white supremacism.

A suspicion of racism

Intel does not present its product as a panacea capable of eradicating hatred online, but as a "not in the right direction, which gives gamers control their experience".The company adds that Bleep, on which it has been working for a few years, has not yet entered the test phase and that many aspects are still called to evolve.

Bleep, le filtre anti-harcèlement d'Intel qui permet de laisser passer «un peu» de racisme

If Bleep can be useful to a young audience that we will prefer to hold as long as possible from this language and moral violence, everything is not going well.

Qualifying the idea of "dystopian", Motherboard notes that this "steps in the right direction" could also constitute a white-only for hate incentors, whose muddy torrents could dump without any restraint, consciousness tranquilized by thisPure product of technological solutionism.

À lire aussiSexisme, cyberharcèlement, entresoi: Twitch confronté à la masculinité toxique

Its strangest characteristic remains the way in which each internet user can filter the remarks made by its congeners and adapt their online experience.

To the way in which we would add without an equalizer a little treble to accommodate its old worn down or bass to energize a part of DOOM, the fine settings of Bleep allow you to choose between several options: to let pass "a little","A large part" or "everything" racist, misogynist, white supremacist, etc..Only the N Word benefits from a clear setting: it is with, or it is without.

Intel indicates that these choices are made on the games machine and not in the cloud, and that a log of "bleeded" remarks, for verification, is made by the software.Which does not solve the fundamental questions posed by the very existence of a tool like Bleep.

How do we quantify "a little" of racism?A white supremacism chouïa?A bit-but-past-trop of validism?Where and on what criteria will the software places the cursors?In addition, will the communities not find, in a few hours, the means and uses to bypass the barriers put in place?