Birdwatch participants can rate others’ grades as a mechanism to prevent users in bad faith from cheating the system and falsely labeling real tweets as fake. These ratings are then merged into a Birdwatch profile separate from a Twitter profile, not unlike the Reddit user rating system.
Twitter said it hopes to build a community of “Birdwatchers” that can eventually help moderate and label tweets in its flagship product, but will not immediately label the tweets with suggestions from Birdwatch.
Twitter has faced increasing pressure over the past year to resolve rampant misinformation on the platform. In addition to the removal, it relied on labeling or adding context below tweets that spread incorrect information. In March, facing a deluge of misinformation about the pandemic, it began to remove “misleading and potentially harmful content” about Covid-19. In May, it launched labels to respond to tweets containing conspiracy theories about the origins of the disease and false cures.
In February, Twitter launched a new “manipulated media” label, first pinning it on a tweet from then President Donald Trump. In the coming months, he would label many more for misinformation about the Covid-19 pandemic and the elections. In the last two weeks before the election, Twitter said he labeled some 300,000 tweets of “disputed and potentially misleading” content.
Twitter told NBC News that it was encouraged by the program’s first tests, which are underway last year. NBC News first reported on a leaked demo of the program, which was then titled “Community Notes”, last February.
Twitter focused heavily on the threat of “manipulation” by what it calls “swarms” of malefactors, who may seek to use the platform as yet another weapon in online information wars.
“We know that there are a number of challenges for building a community-driven system like this – from making it resistant to attempts at manipulation to ensuring that it is not dominated by a simple or biased majority based on the distribution of taxpayers. We will be focused on these things throughout the pilot, ”wrote Coleman.
Researchers will also be able to download bulk data about Birdwatch entries, which he hopes “will allow experts, researchers and the public to analyze or audit Birdwatch” and prevent manipulation.
“We know that this can be confusing and sometimes present problems, but we believe that this model is worth trying,” wrote Coleman.
