Meta, the owner of Facebook and other social media platforms, will implement major changes to its content moderation policies, founder Mark Zuckerberg announced this week in a video titled, “More speech and fewer mistakes.”
Among the changes, Meta’s use of fact-checking organizations will end, and the group will switch to a system of community notes—similar to those used by the X platform—instead.
The recent move comes as tech executives brace for the arrival of incoming U.S. President Donald Trump, whose right-wing supporters have long criticized online content moderation as a tool of censorship.
Why is Meta shifting from fact-checking organizations to community notes?
So why is this happening now, and will it lead to more misinformation? In a video posted to social media platforms, Zuckerberg explained that Meta plans to scrap fact-checking in favor of a new system of community notes, which users can use to identify posts of others that may have misleading or falsified information. Meta plans to roll this community note system out in the next coming months.
Zuckerberg said fact-checking organizations had proved to be “biased” when it came to selecting content to moderate and added that he wanted to ensure free speech on all platforms. “It’s time to get back to our roots around free expression,” he wrote in the post with the five-minute video.
How will community notes impact misinformation on Meta platforms?
“Our system attached real consequences in the form of intrusive labels and reduced distribution. A program intended to inform too often became a tool to censor.” While this policy will extend to all subject matters, Zuckerberg singled out the issues of “gender and immigration” in particular.
Meta’s upcoming modifications will take effect across its trio of major social media platforms: Facebook, Instagram, and Threads, which are used by more than 3 billion people worldwide.
What roles do community notes play in content moderation on Meta?
Presently, social media platforms like Facebook and Threads use third-party fact-checking organizations to verify the authenticity and accuracy of content posted to each platform. These organizations evaluate content and flag misinformation for further scrutiny. When a fact-checker determines a piece of content to be false, Meta will take action to substantially limit that piece of content’s reach, ensuring that it reaches a significantly smaller audience. However, third-party fact-checkers do not have the authority to delete content, suspend accounts, or remove pages from the platform. Only Meta can remove content from its platforms that violates its Community Standards and Ads policies. This includes, but is not limited to, hate speech, fraudulent accounts, and terrorist-related material.
How does Meta’s community notes system compare to X’s approach?
Similar to X, formerly known as Twitter before it was bought for $44 billion by billionaire Elon Musk in 2022, Meta will use community notes to moderate content instead of fact-checkers. X’s community notes, previously known as BirdWatch, was piloted in 2021 and gained significant traction in 2023 as a feature designed to identify and highlight potentially misleading information on the platform. Community notes appear in boxes labelled “Readers added context” below posts on X that have been identified as potentially misleading or inaccurate. A community note typically provides a correction or clarification, frequently supported by a hyperlink to a reputable online source that can verify the information provided.
These annotations are crafted by eligible platform users who have opted into the program. As long as a user has no X violations on their account since January 2023, has a verified phone number provided by a legitimate mobile carrier, and their platform account is at least six months old, they are eligible to participate.
Research supporting the effectiveness of community notes on X
Once approved by X as a notes contributor, participants may rate other community notes as “Helpful” or “Not Helpful.” Contributors receive a “Rating Impact” score that reflects the frequency with which their ratings influence notes that achieve “Helpful” or “Not Helpful” status. A rating impact score of 5 allows a contributor to progress to the next level and write contributor notes for X posts as well as rate them. Community notes that receive five or more ratings undergo algorithmic evaluation. The algorithm categorizes each note as either “Helpful,” “Not Helpful,” or “Needs more ratings.” At this point, the notes are not yet visible to X users, only contributors. Only those notes that receive a final “Helpful” status from the algorithm are shown to all X users beneath the corresponding post.
Are there limitations to the effectiveness of community notes in addressing misinformation?
Although Meta has not outlined exactly how its community notes would work, Zuckerberg stated in his video that they would be similar to X’s community note system. There is debate about how well community notes work on X, which has 600 million users. However, some research into the effectiveness of community notes has been carried out. In October 2024, the University of Illinois published a working paper about X’s community notes feature, led by assistant professor of business administration Yang Gao. In general, his study results were positive.
Another research paper led by University of Luxembourg researchers published in April 2024 in Open Science Framework (OSF)—which allows researchers to share their academic papers—found that the use of community notes reduced the spread of misleading posts by an average of 61.4 percent. However, the research paper added, “Our findings also suggest that community notes might be too slow to intervene in the early (and most viral) stage of the diffusion.”
What reactions have notable figures had regarding Meta content moderation changes?
A recent analysis of notes mentioning election claims between January 1 and August 25, 2024, by The Center for Countering Digital Hate (CCDH) also revealed limits to the effectiveness of the X community notes feature. The researchers examined 283 posts containing election-related claims that independent fact-checking organizations had determined to be false or misleading. In its analysis, it focused exclusively on posts that had received at least one proposed note from community notes contributors. It found that 209 out of 283 “misleading” posts in its sample were not being shown to all X users—“equivalent to 74 percent,” the report stated—because they had not reached the ranking of “helpful” despite being accurate.
Some experts in social media have cautioned that the change may open the door to an increase in misinformation appearing on Meta platforms. Some big names in social media have actively welcomed Meta’s announcement. Elon Musk, CEO of Tesla and who will lead Trump’s Department of Government Efficiency (DOGE) with former Republican presidential hopeful Vivek Ramaswamy, stated, “This is cool” on a recent X post. President-elect Donald Trump appeared to believe he had played a role in Meta’s revised content moderation policy when he spoke at a news conference at Mar-a-Lago on Tuesday. Asked if his previous criticism of the company had prompted the change, Trump’s response was succinct: “Probably.”
Read more: Revamp your scroll | Discover how to reset your Instagram algorithm for a fresh feed
How will the rollout of the new feature proceed across different regions?
Although the initial rollout of the new feature will take place in the next couple of months in the United States only, Zuckerberg also referred to other regions and countries in his video, including Europe, China, and Latin America.
