Commenting on suspicious tweets by volunteers A function to detect the same image will be introduced in 'Community Notes', with comments such as 'This is made by AI' to prevent the spread of fake images



Shortly after

the incident in which the stock market was confused when the fake image 'explosion near the Pentagon' spread to Twitter-certified accounts, a function to detect the same image was introduced on a trial basis to the ' community note ' where users evaluate tweets among themselves. announced by Twitter. As a result, even if multiple accounts spread false information, it may be possible to detect them if they use the same image.

Notes on Media
https://communitynotes.twitter.com/guide/en/contributing/notes-on-media




In the community note, which became officially available in December 2022, users who participate in the community add comments called ``notes'' to tweets. Third-party fact-checking plays a role in preventing misleading or false tweets from being spread as they are.

Here's what the community looks like. There are three tabs, ``Needs evaluation'', ``New note'', and ``Rated as useful'', and each tab is lined with tweets picked up by community participants. The tweet with the large image below was posted by @WarClandestine , who has 238k followers, 'Did you notice that COVID-19 variants stopped after Russia started targeting biological labs? Russia started neutralizing the biological labs in Ukraine on February 24, 2022. No variants since then. Coincidence?' This tweet is open to the public, but only users who have participated in the community note will see the message 'Please help us rate the note sent by the community note contributor'.



By clicking on the message, you can check the notes attached to the tweet by community participants. This user has presented information that `` COVID-19 is still mutating and subspecies are appearing regularly, '' along with multiple sources, and the truth of @WarClandestine's tweet I am questioning. If this note is evaluated as 'useful' by other community participants, the note will be published to general users.



You can also see how Japanese users leave notes on tweets written in Japanese.



On May 31, 2023, Twitter announced that it will introduce a new image detection function to this community note. According to Twitter, users with 10 or more '

Writing Impact ' will see a new option to tag 'About Images' in some Tweets when creating notes. About. This can be selected if the system determines that the image itself may be misleading, regardless of which tweet the image was featured in. Notes tagged 'About images' will now be visible to anyone who mentions an image, and will appear in all tweets that the system recognizes as containing the same image. .



This feature is in the experimental stage at the time of article creation, and is only for tweets with only one image. In the future, we plan to include Tweets with multiple images, as well as Tweets with GIFs or videos.

Related Posts:

in Web Service, Posted by log1p_kr