After violent fighting broke out in Sudan over the weekend, Twitter has been flooded with user accounts spreading false information and propaganda for the warring parties. The subject tag “Sudan” is now trending on Twitter. This means that many tens of thousands of Twitter messages have been written about Sudan in the last few days alone. Many of these messages may have been written by the chatbot ChatGPT, independent fact-checkers show. And much of the content that is spread consists of false information and propaganda. OpenAI, the developer of ChatGPT, warned before launch that their invention could be harmful to humanity. Many now fear that the speech robot makes it easier to spread false information. Photo: LIONEL BONAVENTURE / AFP Thousands of fake messages ChatGPT can easily be used to spread huge amounts of fake information on the internet. – You can use the technology to create lots of user accounts that look real and that produce messages with a certain angle. That’s according to artificial intelligence expert at the Department of Informatics in Bergen, Pål Grønås Drange. – Previously, when you wanted to spread false information, you had to sit down and actually write the false information. It takes time. The more messages, the more time it takes, he says. Therefore, there is a limit to how much false information a person or network can actually spread. But ChatGPT can write tens of thousands of messages in a few seconds, in different languages and as different people, explains Drange. ChatGPT is the best language bot available today. – It does nothing but create text. And it is very good at that, says the researcher. Therefore, it is very difficult to distinguish the text written by ChatGPT from text written by a human. An example of an apparently fake user account on Twitter. This person writes messages with words like “Sudanese” to flood the subject peg with the same name. Photo: Screenshot/Twitter Fake accounts “hijack” subject tags A great many Twitter messages are simultaneously published by user accounts that appear to be fake. At least 900 fake user accounts linked to Sudan have been identified on Twitter. The scope is most likely far greater. Twitter alone has said it deletes up to a million fake accounts every day. Many of the user accounts that have been identified have posted messages with subject words such as “Sudan” or “Khartoum”. This is how you can “hijack” the subject pegs by filling them with empty content, explains Eskil Sivertsen. He is an expert on digital warfare at the Norwegian Defense Research Institute (FFI). One goal of the accounts could be to “drown” real information from Sudan in a sea of fake messages, he explains. – Fake accounts are a huge challenge, says expert on digital warfare at FFI, Eskil Sivertsen. Photo: Norwegian Defense Research Institute This can make it more difficult to find out what is actually happening in the conflict in the country. – It’s about doubting what you see. One sends out large amounts of messages that flood the subject pegs, explains Sivertsen. Such spreading of false information is still nothing new. Among other things, Russia has been very active in sending out false information to manipulate people’s perception of reality. – This is only the tip of the iceberg. Fake accounts are a huge challenge, says Sivertsen. How can you recognize fake accounts on social media? There are several ways to recognize fake social media accounts. First, you can look at the history of the account. When was the account started? What has the account shared before? Who does the account claim to be? On Twitter, all accounts have two names. Is the username the same as the profile name? How many followers/friends does the account have? If an account has no or few followers, this is often a bad sign. Does the account have a profile picture? You can do a reverse image search and see if the image has been shared elsewhere. Losing the battle for the truth The better the technology gets, the more difficult it becomes to distinguish between real and fake information, explains Pål Grønås Drange from the Department of Informatics. – With ChatGPT, those with bad intentions have acquired a very powerful weapon, he says. – How big is the scope of these fake accounts? – We don’t know that. And that is the big problem. There is a race between the social networks that try to limit this and those that spread false information. – Who wins the race? – The technology seems to be winning, time and time again. There is no reason to believe that the information on the internet will be more reliable in the future than it is today, the expert points out, before offering a clear piece of advice: – If you want real information, you have to get it from the people who have the information .
ttn-69