Meta outlines its strategy to combat misinformation ahead of 2022 U.S. midterm elections

Meta has published its strategy for combating misinformation before and during the upcoming 2022 U.S. midterm elections. The policies and techniques to thwart systemic and persistent misinformation campaigns seem quite similar to the company’s strategy back when it was known as Facebook. However, there have been some important additions and revisions.

Meta has confirmed that it will maintain policies and protections “consistent” with the presidential election. Policies barring vote misinformation and linking people to trustworthy information will remain unaltered, just like Twitter. The company will ban political ads. However, the ban will be limited to the last week of the election campaign.

Meta will remove the following data:

• Misinformation about the dates, locations, times, and methods of voting.
• Misinformation about who can vote, whether a vote will be counted, and qualifications for voting.
• Calls for violence related to voting, voter registration, or the administration or outcome of an election.
• Ads encouraging people not to vote or calling into question the legitimacy of the upcoming election.

To fight misinformation, Meta is ensuring better and persistent access to verified and official information from trusted sources. Simply put, the company will be “elevating” post comments from local elections officials in conversations. This should ensure that American voters have easy and ready access to reliable polling information.

Meta indirectly admitted that it had used “Info Labels” excessively during the 2020 elections. The company claimed it will now show such labels in a “targeted and strategic way”.

Apart from sparingly using Info Labels and pushing up official voting information in American voters’ Facebook feed, Meta could also rely on a new AI-powered tool called “Sphere”. The company claims Sphere is “the first [AI] model capable of automatically scanning hundreds of thousands of citations at once to check whether they truly support the corresponding claims.” The AI tool’s dataset includes 134 million public webpages, claimed Meta. The company also has “Few-Shot Learner” to combat misinformation.