The story so far: OpenAI on Thursday (August 16, 2024) said it banned ChatGPT accounts linked to an Iranian influence operation that used the chatbot to generate content to influence the U.S. presidential election.
The Microsoft-backed company said it identified and took down a “cluster of ChatGPT accounts” and that it was monitoring the situation. What is Storm-2035? OpenAI assigned the group the name Storm-2035 moniker, and said the operation was made up of four websites that acted as news organisations. These news sites exploited issues like LGBTQ rights and Israel-Hamas conflict, to target U.
S. voters. The sites also used AI tools to plagiarise stories and capture web traffic, per a Microsoft Threat Analysis Center (MTAC) report issued on August 9.
Some named sites included EvenPolitics, Nio Thinker, Westland Sun, Teorator, and Savannah Time. The operation allegedly targeted both liberal and conservative voters in the U.S.
How did the group use ChatGPT? According to OpenAI, the operatives used ChatGPT to create long-form articles and social media comments that were then posted by several X and Instagram accounts. (For top technology news of the day, subscribe to our tech newsletter Today’s Cache) AI chatbots such as ChatGPT can potentially assist foreign operatives fool gullible internet users by mimicking American users’ language patterns, rehashing already existing comments or propaganda, and cutting down the time it takes to create and circulate plagiarised con.