OpenAI dismantles two influence operations related to Russia and Iran

OpenAI has identified and stopped two disinformation campaigns using ChatGPT to spread false content about geopolitical conflicts. One aimed to denigrate Ukraine, while the other sought to publish fake articles about the conflict between the United States and Iran. OpenAI is strengthening its vigilance against the misuse of artificial intelligence for opinion manipulation. In a recent report, the company reveals that it has neutralized two influence operations named "Dark Clark" and "Bogus Bylines," attributed respectively to Russian and Iranian actors. The first campaign targeted several Latin American countries. According to OpenAI, the operators used ChatGPT to prepare content aimed at denigrating Ukraine and certain Latin American leaders. Among the actions recorded is the distribution on TikTok of a fake audio recording attributed to the Ukrainian consul in Ecuador. The operators also reportedly used a fake email address to incite Peruvian schools to organize events related to Ukraine. The company rates this operation at level 5 out of 6 on its scale for assessing the impact of influence campaigns, while specifying that some of the operators' claims could not be verified. The second operation, attributed to Iranian actors, was based on the publication of fake articles concerning the conflict between Washington and Tehran. ChatGPT was allegedly used to improve the texts, adapt them to the editorial requirements of online media, and prepare submission emails. The content was signed by seven fictional authors posing as Western journalists. OpenAI claims to have spotted nearly 100 articles published in more than a dozen media outlets since July 2025. These revelations illustrate an evolution of disinformation strategies. According to OpenAI, inserting false content into existing media can offer a potentially greater reach than campaigns based solely on fake accounts on social media.