Iranian users used ChatGPT to plant AI-generated articles in US media, report says

Iran-linked operatives used ChatGPT to produce and place articles under fake journalist identities in media outlets worldwide, raising concerns about artificial intelligence being exploited to influence public opinion.

According to The Washington Post, the operation involved an opinion article published by Florida newspaper Port St. Joe Star, which was later identified as AI-generated content linked to an Iranian influence campaign.

OpenAI said on Thursday it had banned accounts associated with the operation after identifying nearly 100 articles published or syndicated across roughly a dozen online outlets. The articles, written in English using prompts in Persian, criticised US military actions against Iran.

The campaign relied on seven fabricated journalist identities, complete with biographies and social media profiles, to make the content appear credible. Some articles also appeared on platforms with substantial online audiences.

OpenAI said the operation appeared to be commercially organised, although it could not identify the actor behind it. The company shared its findings with relevant authorities.