Meta Platforms Inc. plans to make a set of sweeping changes to its fact-checking and content moderation policies, Chief Executive Mark Zuckerberg announced in a video released this morning. 

In 2016, Meta launched a program that relied on third-party fact-checking organizations to review user content on Facebook. The company’s initial partner roster included more than 90 such organizations who reviewed content across 60-plus languages. In 2019, Meta extended the fact-checking program to Instagram. 

As part of the policy changes announced today, the Facebook parent is ending the program. It’s set to be replaced with a Community Notes feature that will rely on Meta’s users to review posts’ accuracy. 

According to the company, Community Notes will allow users to submit notes sharing their views about a post as well as review one another’s notes. Meta says the feature “will require agreement between people with a range of perspectives to help prevent biased ratings.” The company will roll out Community Notes in the U.S. over the next two months, with plans to make it available internationally about a year from now. 

“We’re going to get back to our roots and focus on reducing mistakes, simplifying our policies and restoring free expression on our platforms,” Zuckerberg (pictured) said in the video announcing the changes. 

The second set of policy changes that Meta announced today focuses on political content. For several years, the company limited the visibility of such content in users’ feeds. Going forward, Meta plans to more frequently display political posts from people and Facebook pages that users follow.

Meta’s systems will decide how and when to display political posts based on two main types of data points. According to the company, it will consider explicit signals such as whether a user has liked a political post. Meta also plans to take into account “implicit signals” such as post views. For users who don’t wish to see political content in their feeds, the company will offer settings for filtering such content. 

The third set of policy changes the company announced today focuses on its automated moderation systems. Meta uses them to remove content that breaches its terms of service. The company intends to narrow the systems’ focus to “high-severity violations” such as posts that promote illegal activities. 

When it comes to less severe policy violations, Meta plans to significantly relax its enforcement policies. The company will only take moderation actions in connection with such posts if users report them. Additionally, Meta will require its systems to have “a much higher degree of confidence” that its terms of service have been breached before a piece of content is taken down.

The company is also restructuring its content trust and safety teams as part of the effort. Meta will move the teams from California to “other U.S. locations” including Texas. Additionally, the company said it will more frequently require “multiple reviewers to reach a determination” before removing a piece of content.

In some cases, Meta will use a large language model to provide a second opinion before taking down content. It will also make it simpler for users to appeal policy enforcement decisions. To increase transparency, Meta plans to share statistics about its content moderation mistakes.

According to The Verge, the announcement of the new content policies comes a few weeks after Meta received a letter from incoming Federal Communications Commission chair Brendan Carr. The letter accused the company of participating in a “censorship cartel.” Carr went on to write that Meta could face a government review of its content moderation activities. 

Images: Meta