Meta To Label AI-Generated Content On Facebook "Made with AI" for Transparency

By Thea Felicity

Apr 05, 2024 01:23 PM EDT

This illustration photo taken on February 6, 2023 shows a Facebook ad for Cloud Tech blocked for false information as a phone screen plays a video about the scam. - AFPs fact checkers have recently uncovered a slew of energy-related scams that have proliferated on Facebook from fake incentives for solar panels in the United States to hoax electric bike giveaways in Indonesia and the sale of dud power saving devices in the Philippines. And the trend underscores how fraudsters worldwide profit off disinformation, casting a wide net across social media users, many of whom take the bait amid a cost of living crisis and high utility and energy costs.
(Photo : Photo by JIM WATSON/AFP via Getty Images)

Meta announces new measures to address manipulated media and AI-generated content after feedback from the Oversight Board and extensive policy review processes involving public opinion surveys and expert consultations.

Moving forward, Meta will implement "Made with AI" labels on a wider range of video, audio, and image content when industry-standard AI image indicators are detected or when individuals disclose that they're uploading AI-generated content. 

This decision aligns with the Oversight Board's recommendation to provide transparency and additional context rather than removing content, thereby avoiding unnecessary restrictions on freedom of speech.

Meta also acknowledges that its previous approach was too narrow, focusing primarily on AI-generated videos. With the evolution of technology, especially in the realm of realistic AI-generated content like audio and photos, Meta recognizes the need to address a broader range of manipulated media. 

These changes aim to give users more information and context about the content they encounter on Meta platforms.

READ NEXT: Meta AI Image Generator Branded as 'Racist' After Failing to Generate Images of Interracial Couples

When Will Meta Label AI-generated Content?

Meta plans to begin labeling AI-generated content in May 2024 and will cease removing content solely based on its manipulated video policy in July. This phased approach allows users time to understand the self-disclosure process and adjust to the new labeling system.

The decision-making process behind these changes involved extensive consultations with stakeholders worldwide, including over 120 stakeholders across 34 countries, and public opinion research involving more than 23,000 respondents in 13 countries. 

Most stakeholders supported the implementation of warning labels for AI-generated content depicting people saying things they did not say.

READ MORE: Meta Is Developing a Giant AI Model to Power Its 'Video Ecosystem' Across Its Platforms, Including Facebook 

© 2024 VCPOST, All rights reserved. Do not reproduce without permission.

Join the Conversation

Real Time Analytics