With AI images now going viral across Twitter, the social media platform has decided to update its crowdsourced “fact checking” feature to directly deal with not just the tweet that includes the media, but the media itself.
Twitter is now rolling out(opens in a new tab) “Notes on Media” as part of its Community Notes program. Users who take part in Community Notes, which allows users to add context or fact checks to tweets, will now be able to add these addendums to AI-created photos and other forms of manipulated media as well.
Previously Community Notes users could add notes at the tweet level and the note would follow the tweet around in retweets, quote tweets, and embeds. However, if the actual image itself was reuploaded in a new post, the Community Note on the original upload which added context to the image would not be included. Twitter says the Community Note will now follow(opens in a new tab) the image around as well and will “automatically appear on recent & future matching images.”
Users who want to take part in Community Notes need to apply for the program. Once accepted, they can start rating existing Notes. Once they rate enough Notes that get added to a tweet, their Rating Score will go up and they will then be able to write their own Community Notes. But, in order to write Notes on images, users must have had their own originally authored notes added to a tweet, which builds their Writing Impact Score. A Writing Impact Score of 10 will allow users to write a Community Note on an image.
Notes on Media looks just like the previously existing Community Notes label, except it denotes that “Readers added context to this image.”
This feature should also not be confused with Twitter’s previously existing(opens in a new tab) “Manipulated media” label, which is affixed by the company itself to a tweet containing “synthetic” or “manipulated” images or video. Twitter launched that label back in 2020 in order to combat AI-created images when deepfakes exploded in popularity.
For now, Notes on Media will only work on tweets containing a single still image. Twitter says it hopes to roll out the program for videos, GIFs, and tweets containing multiple pieces of media in the near future.
As the feature was just launched, it remains to be seen how well it’ll work. Twitter will need a system in place to detect reuploads and users have found various ways over the years to avoid content detection on platforms like Facebook and YouTube in order to get around copyright issues.
