The Oversight Board says Meta isn’t doing enough to fight celeb deepfake scams

Scams using celebrity AI buttocks have become an increasingly important question for Meta in the past two years. Now the supervisory board has weighed And apparently confirmed what other criticisms said: Meta is not doing enough to apply her own rules, and makes too easy for the crooks to get away with these programs.

“Meta probably makes it possible to allow large quantities of scam content on its platforms to avoid potentially lighting a small subset of real celebrity mentions,” wrote the board of directors in its last decision. “Examinators on a scale are not allowed to apply this content ban which establishes a false character or pretends to be a famous person to arna -matter or fraud.”

This conclusion came following an affair involving an ad for an online casino style game called Plinko who used a video managed by Ai-Gérée by Ronaldo Nazário, a retired Brazilian football player. The announcement, which, according to the board of directors, has shown obvious signs of being false, has not been deleted by Meta even after being reported as a scam more than 50 times. Meta then deleted the announcement, but not the Facebook publication underlying him until the supervisory board agrees to review the case. It has been seen more than 600,000 times.

The Council indicates that the case highlights fundamental faults in the way in which the meta addresses the moderation of the content for reported scams involving celebrities and public figures. The Council indicates that META has declared to its members that “he applies the policy only on climbing to ensure that the person represented in the content has not really approved the product” and that the interpretation of the individual examiners of what constitutes a “false personality” could vary to the other and introduce incompetences in matters of application “. The result, according to the supervisory board, is that a” significant “quantity of” The meshes of the net.

In its sole recommendation in Meta, the board of directors urged the company to update its internal guidelines, to empower content examiners to identify these scams and to train them on “indicators” of the content managed by AI. In a statement, a Meta spokesman said that “many council affirmations are simply inaccurate” and underlined a test It started last year which uses facial recognition technology to combat frauds “of celebrity baits”.

“Scams have increased on a scale and complexity in recent years, driven by ruthless cross-border criminal networks,” said the spokesperson. “As this activity has become more persistent and sophisticated, so we have our efforts to fight it.

Scams using the depths AI of celebrities have become a major problem for Meta because AI Tech becomes cheaper and more easily accessible. Earlier this year, I pointed out that dozens of pages broadcast advertisements with deepfakes Elon Musk and Fox News personalities promoting supplements claimed to cure diabetes. Some of these pages have repeatedly executed hundreds of versions of these advertisements with apparently little repercussions. Meta has disabled some of the pages after my reports, but announcements of similar scams persist on Facebook to date. Actress Jamie Lee Curtis also recently publicly slammed Mark Zuckerberg for not having withdrawn a deeply fraffe Facebook announcement that presented it (Meta deleted the announcement after its public publications).

The supervisory board also highlighted the extent of the problem in this case, noting that it had found thousands of video announcements promoting the Plinko application in the Meta advertising library. He said several of them presented AI Deepfakes, including advertisements featuring another Brazilian football star, Cristiano Ronaldo, and Meta CEO, Mark Zuckerberg.

The surveillance board is not the only group that has raised the alarm of the scams on the Meta platforms. The Wall Street Journal recently reported That Meta “represented almost half of all the scams reported on Zelle for Jpmorgan Chase between the summers of 2023 and 2024” and that “British and Australian regulators found similar fraud levels from Meta platforms”. The document noted that Meta is “reluctant” to add friction to its process of purchasing advertisements and that the company “reluctant” to prohibit advertisers, even those who have history of scams.

If you buy something via a link in this article, we can win a commission.