Twitter introduces Birdwatch platform for crowdsourcing fact checks

Twitter launches Birdwatch, a community platform that allows users to identify tweets they believe are misleading or inaccurate and provide additional information. It is currently a pilot version.

Users who join the pilot of Birdwatch can add notes to potentially inaccurate or misleading tweets. In it, users can indicate what would be wrong with the tweet and then substantiate this with sources. These Birdwatch annotations can also be reviewed by other users.

The Birdwatch pilot is currently only available in the United States. Various requirements are set for potential participants. Users from the US with a verified email address and phone number can sign up for the pilot, provided they have not previously been banned from the platform for violating Twitter’s rules. Users must also have two-factor authentication enabled.

The pilot will initially consist of 1000 participants and fact checks that result from this will not be visible on Twitter itself in the early stages, the New York Times writes. Instead, the annotations appear on a separate website. If the experiment proves successful, the pilot will expand to more than 100,000 users over the coming months and the Birdwatch contributions will also become visible to all users. Ultimately, Birdwatch should be made available to the “global Twitter audience.” That should happen as soon as there is ‘a consensus of a broad and diverse group of contributors’.

Twitter states that it wants to take several steps to ensure the transparency of Birdwatch. For example, data contributed to Birdwatch must be made publicly available. Users can download this data as tsv files. The code of algorithms that Birdwatch will use will also be made open source. For example, a snippet of the code of the rating system is already online.

How Birdwatch works. Pictures via Twitter