OpenAI thinks it found something larger than a handful of false posts. The corporation terminated a group of accounts linked to Russia that it believed were being used to promote a covert online influence effort.
The operation used ChatGPT to produce social media comments on platforms including Telegram, X, Facebook, and LinkedIn, said OpenAI.
The operators also ordered the algorithm to suppress linguistic indications that could hint at their Russian origins. The infrastructure behind the case makes the case more fascinating.
OpenAI said the campaign promoted a group called the International Burke Institute, which included plagiarized academic material, fake credit and a “sovereignty index” that presented Russia in a positive light while denigrating Western nations.
The company claimed the initial scope of the effort seemed limited. But it was not the major worry. The greater risk was that AI was being used to assist build the illusion of credibility around a broader influence operation, including experts, research and institutional branding.
This makes the episode less about a misinformation campaign, and more about how cheaply standard artificial intelligence can let bad actors generate authority at scale.