Twitter introduces new feature to combat viral fake images created with AI

In recent times, the viralization of false images -mostly created with AI- has become commonplace on social networks. The arrest of Donald Trump, the appearance of a voluptuous (and non-existent) Japanese Minister of Health, or the peculiar Balenciaga-style coat of Pope Francis, have been some of the cases that have attracted the most attention. Thus, Twitter has announced measures to make it easier to identify apocryphal photos and avoid hoaxes.

The platform introduced a new feature of its Community Notes which, henceforth, will allow flag viral images that have been created with AI. This new tool is in the testing phase, but it is expected that it will soon become available to a larger number of users.

Similar to Community Notes, where users can add links and context to a Tweet whose content is deemed false, misleading, or biased, Twitter will offer a similar experience in its new image tool. Although its scope will be even broader. It is that you can check the veracity of a tweet with a fake photo created with AI, or the image itself.

What does the latter mean? That if you choose to mark an image as created with AI, the warning will appear on all tweets that use it, not just the original one. No matter how many times it is republished. This is really important, as it’s quite common for viral content to be shared from scratch across multiple accounts beyond the one that originally posted it.

Twitter wants to make it easier to identify fake photos created with AI

Twitter’s decision to go out to combat the viralization of false photos created with AI comes just days after the image of an alleged explosion at the Pentagon. A story that not only caused public panic — it was retweeted by multiple verified accounts, including one posing as an affiliate of Bloomberg— but also briefly in the US stock market.

For now, the feature to identify fake photos will be available to Community Notes contributors with a score of 10 or higher. Also, initially the feature will be limited in scope, since it will only work with tweets that share a single image.

From Twitter they are already working to expand it to posts with multiple photos and videos. They have also clarified that, in principle, it is likely that alerts will not appear on all images matching those reported as false. However, we will seek to adjust the algorithms to be as accurate as possible.

The hard fight against misinformation

The circulation of fake news on Twitter is old history, but the problem seems to have worsened since Elon Musk bought the social network. In fact, Thierry Bretonthe European Commissioner for the Internal Market and Services, pointed out in recent days against the American company.

“Twitter abandons the European Union’s voluntary Code of Practice against disinformation, but the obligations remain. You can run, but you can’t hide. Beyond the voluntary commitments, combating disinformation will be a legal obligation under the Digital Services Act to starting August 25. Our teams will be ready for your application”, tweeted the official.

In any case, it is also important to note that the problem of the viralization of false images goes beyond Twitter. Let’s not forget that Midjourney, one of the most popular generative AI platforms, was forced to cancel its free trials over controversy with Trump and Pope Francis.