Twitter is planning to expand its crowdsourced fact-checks program and bring it to images for its users in the United States. This comes after fake images of an explosion in the Pentagon went viral on the micro-blogging platform.
User-generated Community notes that appear below a tweet to inform of potentially misleading information will now also appear for images. Now, contributors will be able to add information specifically for images on Twitter and that context will populate below recent and future matching images.
In its tweet, Community Notes specifically mentioned AI-generated images, pressing on the recent development in the technology, and its ability to create believable realistic images.
Twitter introduced the Birdwatch program (now Community Notes) back in 2022, as “a pilot in the US of a new community-driven approach to help address misleading information on Twitter.” Over time, the Birdwatch program has gone through new developments.
In the starting phase, notes were only visible on a separate Birdwatch website, where participants could also rate the helpfulness of notes added by other contributors. The functioning of this system relies heavily on Birdwatch’s bridging algorithm, which sets it apart from conventional Twitter algorithms.
Unlike algorithms that prioritize content based on majority consensus or rank it higher when it achieves a certain level of engagement, Birdwatch takes a unique approach. Its algorithm aims to identify consensus across diverse groups that hold contrasting viewpoints, and only then does it bring attention to the fact-checks contributed by the crowd, making them visible to other Twitter users.
To qualify as a Birdwatch contributor, users are required to demonstrate their ability to write informative annotations, which provide additional context to tweets. Twitter evaluates potential contributors based on their “rating impact” score. This score initially starts at zero and needs to reach a minimum of 5 for an individual to become a Birdwatch contributor. Users accumulate points by rating Birdwatch notes that contribute to the designation of “Helpful” or “Not Helpful” for those notes. Conversely, they lose points when their ratings contradict the final status assigned to the note.
Recently, Twitter has been at the receiving end of the controversy regarding misinformation on the platform. Not too long ago, the platform granted blue checks to deceased celebrities creating a very awkward situation. However, the new attempts seem positive and appreciative.
It was only last month when the company decided to expand Birdwatch right before the mid-term elections in the United States. Now, Twitter plans to bring crowdsourced fact checks to Images to push the community efforts to eliminate fake information from the platform.
Source: Community Notes Tweet