Social media: International authorities smash Russian AI bot farm
Russia's troll factories use AI to create and control fake social media profiles. One was shut down.
(Image: Bild erstellt mit KI in Bing Designer durch heise online / dmk)
US authorities around the Department of Justice (DoJ) have taken a Russian state bot farm for social media offline. It used an AI-based tool to exert malicious influence on other countries with disinformation campaigns.
The Department of Justice announced that the attack was successful as part of an international operation. Two domain names were confiscated, which served as a control center for bot farms, and 968 X accounts (formerly Twitter) were searched. These were used by Russian actors to set up an AI-supported social media bot farm and spread disinformation in the USA and other countries.
Fake accounts with supposedly local connections
The social media bot farm used artificial intelligence to create fictitious social media profiles. These often pretended to belong to people in the United States or the other countries under attack. The operators used these fake profiles to spread messages in support of Russian government targets.
Together with international law enforcement agencies from Canada and the Netherlands, for example, the FBI has published an analysis of the "Meliorator" control software used (PDF). According to the analysis, actors associated with RT (formerly Russia Today), a Russian state-owned media organization, used Meliorator to exert malicious influence abroad with a focus on the goals of the Russian government. The activities took place in Germany, Israel, the Netherlands, Poland, Spain, Ukraine and the United States.
So far, the use of Meliorator on X in particular has been detected. However, the investigating authorities have found evidence that the developers also wanted to extend the functions to other social media platforms. Meliorator has extensive functions:
- Mass creation of authentic-looking social media profiles,
- Disseminating content similar to that typically created on social networks,
- Mirroring disinformation from other bot profiles,
- Repeating existing false narratives to reinforce malign foreign influence, and
- Formulating messages, with content and framing, based on the bot archetype.
It is therefore a powerful and comprehensive control tool. The prosecutors explain the individual functions, some of which were supported by AI, in more detail in the PDF. At the end, the authors of the analysis summarize tips and advice that social network operators can take as countermeasures against the infiltration of Russian state actors and their disinformation campaigns.
Disinformation campaigns are currently omnipresent on all social networks. Facebook, for example, is currently experiencing a flood of "Old times were good" or "Grandma's best recipes" profiles that aim to stir up emotions with old stories - including bizarre ones such as "The GDR was so beautiful". The aim is to polarize and divide society. Targeted Russian disinformation with a wide reach on social networks has been a problem for some time. In 2019, for example, US authorities cut off a Russian troll factory from the internet ahead of the US election.
(dmk)