OpenAI’s recent discovery of an Iranian influence operation utilizing ChatGPT has raised concerns about potential interference in the upcoming U.S. presidential election. The group, known as Storm-1679, was found to be generating articles and social media content focused on shaping public opinion around Vice President Kamala Harris, former President Donald Trump, and various other political topics.
Despite OpenAI’s findings, the impact of Storm-1679’s efforts appears to have been minimal, with most posts and articles receiving little engagement from real users. This particular operation targeted a range of issues, including the 2024 presidential candidates, Israel’s invasion of Gaza, U.S.-based Latinx communities, Venezuelan politics, and Scottish independence.
OpenAI took action by banning a cluster of accounts responsible for creating the AI-generated content and sharing threat intelligence with relevant stakeholders. This discovery marks the first time that OpenAI has disclosed a specific election interference campaign involving ChatGPT.
This incident is not an isolated one, as similar reports of online foreign interference in U.S. elections have become more frequent. A recent report from Microsoft highlighted a similar Iranian network, Storm-2035, which created polarizing content related to the election, LGBTQIA+ rights, and Israel’s actions in Gaza.
Foreign attempts to influence U.S. elections through various means, such as phishing attacks and misinformation campaigns, continue to pose a threat to the democratic process. It is crucial for technology companies, government agencies, and political campaigns to remain vigilant and take proactive measures to combat these threats.
