Twitter unveiled a feature Monday meant to bolster its efforts to combat misinformation and disinformation by tapping users in a fashion similar to Wikipedia to flag potentially misleading tweets.
The new system allows users to discuss and provide context to tweets they believe are misleading or false. The project, titled Birdwatch, is a standalone section of Twitter that will at first only be available to a small set of users, largely on a first-come, first-served basis. Priority will not be provided to high-profile people or traditional fact-checkers, but users will have to use an account tied to a real phone number and email address.
“Birdwatch allows people to identify information in Tweets they believe is misleading or false, and write notes that provide informative context," Twitter Vice President of Product Keith Coleman wrote in a press release. "We believe this approach has the potential to respond quickly when misleading information spreads, adding context that people trust and find valuable."
While Birdwatch will initially be cordoned off to a separate section of Twitter, the company said “eventually we aim to make notes visible directly on Tweets for the global Twitter audience, when there is consensus from a broad and diverse set of contributors.”
Demos of the product viewed by NBC News showed a separate area in which tweets are discussed and rated in a format that combines elements of both Reddit’s and Wikipedia’s moderation tools.
Birdwatch users are able to flag tweets from a dropdown menu directly within Twitter’s main interface, but discussion about a tweet’s veracity will remain exclusively in the Birdwatch section. Twitter says it does anticipate some users linking directly to Birdwatch discussions underneath high-profile and controversial tweets, just as some users would link out to fact-checking sites.
Participants in Birdwatch are able to rate others’ notes, as a mechanism to prevent bad-faith users from gaming the system and falsely labeling true tweets as false. Those ratings are then assembled into a Birdwatch profile separate of a Twitter profile, not unlike Reddit’s user-rating system.
Twitter said it hopes to build a community of "Birdwatchers" that can eventually help moderate and label tweets in its main product, but will not be immediately labeling tweets with Birdwatch suggestions.
Twitter has faced increased pressure over the last year to address rampant misinformation on the platform. Aside from removal, it has relied on labeling, or adding context below tweets that spread misinformation. In March, facing a deluge of misinformation about the pandemic, it began removing “misleading and potentially harmful content” about Covid-19. By May, it had introduced labels to respond to tweets containing conspiracy theories about the origins of the disease and fake cures.
In February, Twitter rolled out a new “manipulated media” label, affixing it first to a tweet from then-President Donald Trump. In the months ahead, it would label many more for misinformation around the Covid-19 pandemic and the election. In just the final two weeks before the election, Twitter said it labeled some 300,000 tweets for “disputed and potentially misleading” content.
Twitter told NBC News it was encouraged by early trials of the program, which have been ongoing in the last year. NBC News first reported on a leaked demo of the program, which was then titled “Community Notes,” in last February.
Twitter heavily focused on the threat of “manipulation” by what it calls “swarms” of bad actors, who may seek to use the platform as another weapon in online information wars.
“We know there are a number of challenges toward building a community-driven system like this — from making it resistant to manipulation attempts to ensuring it isn’t dominated by a simple majority or biased based on its distribution of contributors. We’ll be focused on these things throughout the pilot,” Coleman wrote.
Researchers will also be able to download bulk data about Birdwatch entries, which he hopes will “enable experts, researchers, and the public to analyze or audit Birdwatch” and deter manipulation.
“We know this might be messy and have problems at times, but we believe this is a model worth trying,” Coleman wrote.