OpenAI said on Friday it discovered an “Iranian influence operation” using ChatGPT. The group, known as Storm-1679, published articles and social media comments to shape public opinion around Vice President Kamala Harris and former President Donald Trump, according to OpenAI. OpenAI said that in addition to targeting 2024 U.S. presidential candidates, Storm-1679 also generated content about Israel’s invasion of Gaza and its participation in the 2024 Olympics, the rights of the Latino community in the United States, Venezuelan politics, and Scottish independence from the United Kingdom.
The company said most posts and articles discovered by OpenAI had little or no approval from real users. Still, it detailed the incident on its blog, writing that it had discovered more than a dozen X (formerly Twitter) accounts posing as conservatives and progressives, using the words “#DumpTrump” and ” #DumpKamala” and other tags. According to OpenAI reports, Storm-1679 also used at least one Instagram account to spread artificial intelligence-generated content.
OpenAI has previously profiled “state-related threat actors” using its tools, but this is the first time it has disclosed a specific election interference campaign utilizing ChatGPT.
OpenAI said it responded to the findings by banning the “cluster” of accounts that created the content; the company also said it “shares threat intelligence with government, campaign and industry stakeholders.” The company did not identify those interests. name of the author, but it did share some screenshots of the post. The screenshots have views ranging from 8 to 207, with barely any likes.
Storm-1679 also shared ChatGPT-generated articles on multiple websites, which “impersonated both progressive and conservative news outlets,” OpenAI said. The company added: “Most of the social media posts we found received few or no likes, shares or comments. We also found no indication that online articles were shared on social media.
An August 6 report from Microsoft similarly described Storm-2035, an Iranian network that hosts “four websites disguised as news outlets.” According to Microsoft, the network produced “polarizing” posts about the election, LGBTQIA+ rights and Israel’s invasion of Gaza.
Reports of foreign cyber interference in U.S. elections are now almost commonplace. For example, Microsoft’s Aug. 6 report also detailed an Iran-linked phishing attack that targeted an unnamed “senior” U.S. campaign official. Shortly after Microsoft dropped the report, the Trump campaign announced that “foreign sources” had stolen some of its emails and documents in an attempt to influence the 2024 presidential election. Eight years ago, a Russia-linked hacking group called Guccifer 2.0 stole Democratic National Committee emails through similar phishing attacks; they eventually leaked thousands of Democratic emails ahead of the 2016 Democratic National Convention National Committee emails and documents.
Under pressure from lawmakers, Big Tech has rolled out various measures over the years to address such incidents. Their efforts included meme fact-checking, wishful thinking, a short-lived political ad ban, “war rooms” and collaboration with rivals and police.
