Twitter to expand its community fact-checking experiment, Birdwatch

Twitter will expand community fact-checking project called Birdwatch as an approach to new form of content moderation

Twitter Inc will expand its community fact-checking project called Birdwatch, the social media company said on Wednesday, deepening its novel approach to a new form of content moderation.

Birdwatch launched last year, allowing some Twitter users to debunk misleading tweets by attaching notes to the content in order to provide context or point to accurate sources.

Social media platforms including Twitter have long faced competing pressures on how to moderate content that appears on their services. Critics have accused the companies of doing too little to remove harmful posts, while others argue the platforms should protect free speech.

Billionaire Elon Musk, who is attempting to walk away from his $44 billion agreement to buy Twitter, has said the company should remove fewer posts and act as a public town hall for free speech.

Tweets with Birdwatch notes are left up on the service and its algorithmic distribution to other users is not affected.

"We just think that's a really powerful place to start, because it's just arming people with information and letting them make up their own minds," said Keith Coleman, vice president of product, during a briefing with reporters.

While Twitter has policies that prohibit content such as hate speech or calls for violence, Birdwatch allows the Twitter community to address tweets in "gray areas," he said.

Until now, Birdwatch was a limited experiment with 15,000 contributors writing fact-checking notes. Twitter said it will now add about 1,000 new contributors per week.

Birdwatch notes are held on a separate website, but half of users in the United States will begin seeing notes in their Twitter timelines, the company said.

The project has produced encouraging results, Twitter said. People are 15% to 35% less likely to "like" or retweet content that has a Birdwatch note attached to it. They are also 20% to 40% less likely to agree with a potentially misleading tweet after reading a Birdwatch note about it.