Previously, social media platform X blocked part of the search for Taylor Swift because deeply forged (deepfake) images of singer Taylor Swift (Taylor Swift) circulated online.
On Monday, when searching the site for her English name, without quotation marks, a message appeared that showed an error and prompted the user to retry the search, with the sentence: "Don't worry, it's not your fault."
However, if you enclose her name in quotation marks, posts that mention her name can still appear.
Last week, colored and malicious images related to Taylor Swift began to spread widely on X, and technology platforms and anti-abuse groups have been working to solve the problem.
"this is an interim measure and we are very cautious in dealing with this issue because we put safety first," Joe Benarroch, head of business operations at X, said in a statement.
Unlike the traditional fake images that have plagued celebrities in the past, these images of Swift seem to have been created using an artificial intelligence image generator that can instantly create new images based on written prompts.
After the pictures began to spread online, the singer's loyal fans of Swift quickly mobilized and fought back on X, using the hashtag # protect Taylor Swift (# ProtectTaylorSwift) to post more positive pictures of the pop singer. Some people say they are reporting accounts that post deeply fake pictures.
Reality Defender, a group that detects deep forgery, says it has tracked down a large number of unapproved colored materials depicting Swift, especially on X. Some of the images have also made it to Meta's Facebook and other social media platforms.
The researchers found at least dozens of unique images generated by artificial intelligence. Among the most widely disseminated were football-related, images of Swift painted or bloodied, objectifying her and, in some cases, inflicting violent harm on her in order to deeply fake her image.
Ben Decker of the threat intelligence group Memetica said that the deeply forged images of Taylor Swift first appeared on marginal social platforms last year during an ongoing campaign to produce indecent images of female celebrities generated by artificial intelligence, and a widely circulated Swift image appeared online as early as Jan. 6.
X's move to reduce search for Swift may just be a stopgap measure.
"When you're not sure what everything is, and you can't guarantee that everything will be deleted, the easiest thing you can do is limit people's ability to search for them," he said.
X lifted temporary measures to block the search for Taylor Swift's name on Monday night, Reuters reported.
Some researchers have pointed out that the number of obvious deep forgery of indecent photos has increased in the past few years as the techniques used to generate such images have become easier to obtain and use.
In 2019, a report released by artificial intelligence company DeepTrace Labs showed that most of the images were aimed at women, and most of the victims were Hollywood actors and South Korean K-pop singers.
In the European Union, separate pieces of new legislation include content aimed at deep forgery. The Digital Services Act (Digital Services Act), which came into force last year, requires online platforms to take measures to curb the risk of spreading content that infringes on "fundamental rights" such as privacy, such as "unauthorized" images or deeply forged pornographic content.
The EU is awaiting final approval of the artificial Intelligence Act (Artificial Intelligence Act), which will require companies that use artificial intelligence systems to create deep forgery to also tell users whether their content is artificial or forged.