Unlike Polis, Twitter’s project has to find consensus not for one question at a time but for any conceivable controversy on the platform. Community Notes does that by estimating the diversity of viewpoints among participants, based on how different users rate the helpfulness of notes from others. Twitter’s version of bridging the divide is to find notes considered useful by raters who don’t usually agree on much, suggesting they have different viewpoints. This technique at the heart of Polis and Community Notes is called bridging-based ranking. Megill believes Twitter’s team has taken it to new technical heights. “Birdwatch has made a profound breakthrough in scaling this kind of system across a bigger population and number of issues,” he says.
Twitter’s vice president of product said earlier this month that usage of Community Watch has recently spiked, but the project is still in its early days. The data is open source, and as of November 8 it had seen only 38,494 notes from 5,433 people—a small group to oversee a platform with more than 200 million users. Nor can bridging-based ranking change human nature. One independent study found that people are more likely to write notes on tweets expressing viewpoints different from their own. David Rand, one of its authors, concluded in the Financial Times that “partisanship is a major driver of users’ engagement on Birdwatch.”
Twitter’s own recently released research also reports a partisan divide, with many more Democrats than Republicans finding the notes helpful. But a majority of both groups thought the notes selected by the system were helpful rather than not helpful. And Community Notes were also seen to reduce how much users share Tweets shadowed by heavily caveating notes. The project can also claim some notable, if anecdotal, victories: This month both the White House and Elon Musk deleted widely circulated tweets after a Community Watch note called out missing context.
Perhaps Community Notes’ biggest weakness is also one shared by Polis. “Those digital democracy platforms don’t have any kind of real authority,” Taiwanese parliamentarian Karen Yu told me. Polis still relies on politicians to turn the consensus it draws from citizens into law. Because the users of a social platform have so little power over the service they use, Community Notes is even weaker. With a flick of his wrist, Elon Musk could make it—and all the community’s notes—vanish.
But I don’t think he will. An old joke about Twitter attributed to Mark Zuckerberg says the company’s management was so clueless that “they drove a clown car into a gold mine and fell in.” Elon Musk may have driven his own clown car into his own gold mine. He seems unlikely to have known that Birdwatch existed before buying the platform, but he has stumbled upon one of the most exciting content moderation innovations ever to come out of not just Twitter, but any major platform.
For Musk, who has loaded Twitter with debt, there is much to love in Community Notes. It is scalable, powered by algorithms, and does not require employing legions of content moderators. Most of all, it transfers responsibility for defining the truth away from Twitter itself and onto its users.