The Meta Ai AI Supervision Council launches “unjustified and unjustified”

Photo of author

By [email protected]


Since Meta platforms fill with more content created from artificial intelligence, the company still has a lot of work that must be done when it comes to applying its policies on the processed media. The Supervisory Council once again criticizes the social media company because of its dealings with these jobs, as it writes in its recent decision that its inability to impose its rules constantly “is undisciplined and unjustified.”

If this seems familiar, it’s because this Since last year, the Supervision Council has used the word “unsessed” to describe a deadly approach to the treated media. The Board of Directors had previously urged Meta to update its bases after a video that was misleading in a misleading manner on Facebook. In response, Mita said that Its use of stickers to determine the content created by artificial intelligence and that it will apply more prominent stickers in “high -risk” cases. Notice these stickers, like those below, when creating or editing a post using artificial intelligence.

An example of the name when Meta determines a piece of content that is manipulated An example of the name when Meta determines a piece of content that is manipulated

An example of the sticker when Meta determines part of the AI ​​manipulated content is “high risk”. (Screen shot (Meta))

The Board of Directors said this approach is still short though. She said in its latest decision: “The Board of Directors is concerned that although the spread of processed content is increased through coordination, a dead enforcement of the processing media policy is not consistent.” “A dead failure to apply the poster automatically to all cases of the same treated media is not coherent and unjustified.”

The statement came in a decision related to the post To display a voice for politicians in Iraqi Kurdistan. The supposed “registered conversation” included a discussion about the forgery of the upcoming elections and other “evil plans” of the region. The council said that this post was reported to Meta to mislead the information, but the company closed the case “without a human review.” Meta was later described as some of the audio clip counterparts, but not that originally reported.

The case, according to the council, is not strange. It seems that Meta told the painting that he could not define stickers and apply them automatically to audio and video posts, only to “fixed images”. This means that multiple counterparts of the same audio or video clip may not get the same treatment, which can be caused by the painting more confusion. The Supervision Council also criticized Meta for its adoption often on the third parties to determine the video and sound concerned with the spontaneous organization, as it did in this case.

“Given that Meta is one of the leading technology companies and artificial intelligence companies in the world, with its resources and the wide use of Meta platforms, the Board of Directors repeats that Meta should give priority to investing in technology to determine and wide video and sound video mark,” Board of Directors wrote. “It is not clear to the Board of Directors the reason for the existence of a company from this technical experience and resources that determine the potential treatment of the media in the positions at risk to the media or reliable partners.”

In its META recommendations, the council said that the company must adopt a “clear process” to place a “identical or similar” content “constantly” in situations when a “high -risk” sign adds to a publication. The Board of Directors also recommended that these posters appear in a language that matches the rest of its settings on Facebook, Instagram and Threads.

Meta did not respond to a request for comment. The company has 60 days to respond to the council’s recommendations.



https://s.yimg.com/ny/api/res/1.2/rJ3Vg0hntRZ4_bLCv_hqhw–/YXBwaWQ9aGlnaGxhbmRlcjt3PTEyMDA7aD04MDA-/https://s.yimg.com/os/creatr-uploaded-images/2025-01/3b4bf6d0-d842-11ef-bbbf-f9cbd8bfcfef

Source link

Leave a Comment