Twitter announces the launch of Birdwatch, a platform to detect “misleading tweets” with the collaboration of the community

Twitter has been going since the final stretch of the US presidential campaign. labeling “potentially misleading” tweets in order to prevent the spread of ‘fake news’. But now the social network wants to go one step further.

Twitter announced a few hours ago the launch of a platform called Birdwatch, with which the company seeks to turn to the user community to deal with ‘misinformation’.

A pilot project called to ‘provide context’ to our tweets

If you try to access, you will come across a notice that the website is only available in the US:

This is because it is a pilot test, but in the future will allow users to register to report possible misleading tweets and add annotations for them to be examined by other collaborators, who will be able to assess their quality.

That it is in testing also explains why Twitter will only make it available, initially, to a small group of users. When it comes to filtering them No priority will be given to “high profile” users or traditional ‘factcheckers’, but it will be a requirement for users that their accounts are linked to real phone numbers.

The goal of Twitter is to help create an ecosystem capable of providing context to tweets and thus, over time, community annotations themselves end up appearing appended to tweets on

According to NBC, the product demos they had access to showed a web space dedicated to the debate on the most controversial tweets, with “a format that combines elements of the Reddit and Wikipedia moderation tools“.

According to Keith Coleman, Twitter’s vice president of product, this approach has garnered support from “individuals across the political spectrum” and has

“Potential to respond quickly when misleading information is spread, by adding context that people can trust.”

Finally, in order to equip “transparency” to this new tool, from Twitter they clarify that they will take two measures:

  • That all data provided is publicly available and are downloadable in TSV format.

  • That the classification criteria of tweets, as well as the algorithm code, are documented and accessible in their own Github space.