Overview
- OpenAI has removed ChatGPT accounts likely linked to Russia.
- These accounts were promoting a fictitious Israeli institute that published plagiarized research under various academic names.
- While most posts attracted minimal attention, their affiliated Telegram channels boasted up to 20,000 followers.
A Russian-origin network has been found to use ChatGPT to endorse an organization masquerading as an academic institute, according to a report by OpenAI released on Wednesday.
The organization stated that it had banned several accounts that utilized Russian-language prompts to produce social media content primarily in English. Additionally, these accounts directed ChatGPT to obscure any indications of their Russian origin.
Myriad: When will OpenAI release GPT-6? Click to make your prediction.OpenAI noted, “Although the campaign seems to have reached relatively small audiences, its intricate design sets it apart from other Russia-linked influence efforts we have countered since the onset of the Ukraine conflict.”
Due to ChatGPT's unavailability in Russia, the group reportedly employed virtual private networks (VPNs) to mask their locations and gain access to the service.
The network promoted the International Burke Institute, which claimed to be an expert group based in Israel, utilizing official-looking profiles and accounts pretending to be regular users. OpenAI discovered that 34 out of 36 articles reviewed on the institute's website were plagiarized, with some incorrectly attributed to academics who had no involvement in the writing.
OpenAI stated, “Some of these articles were dated, and others were assigned to the wrong authors.” For instance, an article about the China-Pakistan Economic Corridor was found to have been copied from a Cambridge University Press original but misattributed to a University of Nottingham professor whose specialization is in South Asian politics.
Other political influence campaigns have similarly utilized generative AI technologies.
In 2024, OpenAI identified another Russia-linked initiative called Doppelganger, which employed ChatGPT to craft articles, social media content, and comments that praised Russia while disparaging Ukraine, NATO, and the United States.
Additionally, groups associated with Russia have leveraged ChatGPT for cyber activities. OpenAI mentioned that Forest Blizzard, a group connected to Russian military intelligence, used the chatbot for target research, scriptwriting, and aiding in phishing attempts.
While the campaign may have involved Russian language prompts, OpenAI did not link the IBI initiative directly to the Russian government.
Myriad: How many days will Claude go down? Click to make your prediction.OpenAI indicated that some individuals in Israel might have represented the institute at conferences or through article submissions, but the nature of their connection to the campaign is still uncertain.
OpenAI concluded, “We cannot ascertain the relationship between these individuals, the IBI, and the Russian operators.”