An Iranian group has been discovered using ChatGPT to generate content aimed at influencing the US presidential election.

Posing as news publishers, "Storm-2035" had operated through four websites - EvenPolitics, Nio Thinker, Westland Sun, and Savannah Time - and used ChatGPT to generate long-form articles and social media comments on various topics, including US presidential candidates, the Gaza conflict, and Israel's participation in the 2024 Olympic Games.

According to OpenAI, which said it had identified and dismantled the network, the content targeted both liberal and conservative voters and exploited divisive issues such as LGBTQ rights and the Israel-Hamas conflict.

OpenAI claims the operation did not achieve significant audience engagement, with most posts receiving little to no interaction, and that it had banned the accounts associated with the Iranian group.

The incident highlights the potential misuse of AI chatbots for creating seemingly credible political propaganda, even if in this instance it appeared to have little impact. It also underscored the importance of properly safeguarding generative AI systems  against potential misuse in political influence campaigns.