OpenAI says Iran tried to influence US elections with ChatGPT

OpenAI has banned a string of ChatGPT accounts tied to an Iranian influence campaign that generated and shared content related to the US presidential election, among other topics. The operation mainly used ChatGPT to create longform articles and social media comments for platforms like Instagram and X, according to OpenAI.

OpenAI linked the accounts to Storm-2035, a covert Iranian influence operation that has attempted to engage US voters by launching websites disguised as political news outlets. In addition to commentary about the US election on both sides of the political spectrum, OpenAI says the operation generated content about the Israel-Hamas war, Israel at the Olympic Games, politics in Venezuela, and “the rights of Latinx communities” in the US.

OpenAI says that its investigation found the campaign “does not appear to have achieved meaningful audience engagement.” The AI company says most social media posts it tracked down received “few or no likes, shares, or comments.” In May, OpenAI and Meta announced they’d disrupted a social media campaign that used AI to post pro-Israel messages on Instagram and Facebook.

With the US presidential election just months away, we may see more attempts to interfere with its outcome. Last week, former President Donald Trump confirmed his campaign was hacked and linked the incident to a phishing email sent by an Iranian hacking group. In June, the FBI opened an investigation into the purported hack on Trump’s campaign as well as alleged hacking attempts on the Biden-Harris campaign.