Este contenido aún no está disponible en Español (España)
AI-Manipulated Video Promoting Gambling
ÚLTIMA ACTUALIZACIÓN 4 AGO. 2025
2025-018-FB-UA
Today, February 6, 2025, the Oversight Board selected a case appealed by a Facebook user regarding a video which depicted an AI-manipulated version of a person appearing to be Brazilian soccer player Ronaldo Nazário encouraging others to download a popular online game called Plinco. The video shows AI-generated images of a schoolteacher, bus driver, and grocery store worker with their average salaries in Brazil, while an AI-generated voice imitating Ronaldo claims the game is easy to play and can earn players more money than these jobs. The video also encourages users to click a link to download Plinco, however, the link leads to a different game.
Upon initial review, Meta left this content up. However, upon further review, we determined the content did in fact violate our Fraud, Scams and Deceptive Practices policy, as laid out in the Facebook Community Standards, and was left up in error. We therefore removed the content.
Under our Fraud, Scams, and Deceptive Practices policy, Meta prohibits content that “attempts to establish a fake persona or to pretend to be a famous person in an attempt to scam or defraud.” By making it appear as if Ronaldo is promoting Plinco through AI, we determined the video attempts to scam people into using a product they might not otherwise download without his endorsement. Under our Spam policy, Meta also prohibits content “containing a link that promises one type of content but delivers something substantially different.”
We will implement the Board’s decision once it has finished deliberating, and will update this post accordingly. Please see the Board’s website for the decision when they issue it.
Read the board’s case selection summary
Case decision
On May 22, 2025, the Oversight Board overturned Meta's original decision to leave the content up. Since Meta previously removed the post, there will be no further action on the case content.
When it is technically and operationally possible to do so, we will also take action on content that is identical and in the same context as the first case. For more information, please see our Newsroom post about how we implement the Board’s decisions.
After conducting a review of the recommendations provided by the Board, we will update this post with initial responses to those recommendations.
Recommendations
Recommendation 1 (Assessing Feasibility)
To better combat misleading manipulated celebrity endorsements, Meta should enforce at scale its Fraud, Scams and Deceptive Practices policy prohibition on content that “attempts to establish a fake persona or to pretend to be a famous person in an attempt to scam or defraud” by providing reviewers with indicators to identify this content. This could include, for example, the presence of media manipulation watermarks and metadata, or clear factors such as video-audio mismatch.
The Board will consider this recommendation implemented when both the public-facing and private internal guidelines are updated to reflect this change.
Commitment Statement: We are assessing the feasibility of implementing a number of updates to our Fraud, Scams, and Deceptive Practices Policy that we believe will address the spirit of this recommendation. These updates include work to build tools that, combined with regional context and existing enforcement approaches, help us address “celeb-bait” content.
Considerations: What constitutes a ‘fake persona’ may differ across regions and identifying celeb-bait and celebrities in general can be difficult at scale. As such, we are working with our engineering and product teams on tools that human reviewers can also work with to identify and face match harmful celeb-bait content at scale. Given the complexity of this work, we continuously work to improve the tools used to detect and prevent scams.
Our Fraud, Scams, and Deceptive Practices Community Standard explains that we aim to protect users and businesses from potentially being deceived out of money, property, or personal information by removing content and addressing behavior that purposefully employs deceptive means. The means described here may include misrepresentation, making exaggerated claims, or using stolen information. As part of our existing approach to this content, we remove content that attempts to establish a fake persona or pretend to be a famous person in order to scam or defraud. This includes removal of manipulated audio and visual content with these intentions. When any content is reported, it is assessed across other Community Standards or policies as well and we may remove or apply a label accordingly if manipulated content violates other policies such as Bullying and Harassment or Misinformation.
We currently enforce this policy related to fake personas on escalation, and we are assessing ways to scale how we enforce this content more broadly. As the Board notes in its decision, assessing if content is ‘fake’ or includes a ‘celebrity’ is a highly context-dependent assessment that can require regional expertise.
We will provide regular updates to the Board in upcoming reports on the status of implementing this recommendation.