OpenAI blocks Iranian group's ChatGPT accounts over US election manipulation

OpenAI has identified and shut down a covert Iranian influence operation that exploited ChatGPT, marking the second disclosure of its AI models being used for adversarial purposes since May.

OpenAI, supported by Microsoft Corp., uncovered the operation, known as Storm-2035, which used ChatGPT accounts to create content aimed at swaying public opinion during the 2024 U.S. elections, Caliber.Az reports via foreign media.

The affected accounts have been banned from accessing OpenAI’s services.

This operation generated long-form articles and short social media posts on various subjects, including U.S. politics and the presidential election, and distributed them through social media and fake news websites. Despite their efforts, the content received limited engagement, with most posts attracting minimal interaction and articles failing to gain significant traction.

OpenAI’s investigation, aided by Microsoft, revealed that the operation primarily focused on topics such as the Gaza conflict, Israel's Olympic presence, and U.S. presidential race developments. OpenAI has shared its findings with government, campaign, and industry stakeholders to help combat similar threats.

This case highlights a growing trend of using AI tools for disinformation campaigns. In May, OpenAI reported that its models were being misused by networks from Russia, China, Iran, and Israel to spread false information globally. The company had previously disclosed five such covert influence operations using its AI.

In response to rising geopolitical tensions, OpenAI announced plans in June to restrict access to its tools in China. However, some developers in China have continued to use these tools through VPNs and other methods.

Election security concerns have escalated following a cyberattack on Donald Trump’s campaign, with warnings about potential future cyber threats echoing concerns from previous election cycles.

On August 15, Google confirmed that Iranian hackers linked to the Revolutionary Guard had targeted the personal email accounts of individuals connected to the U.S. presidential campaigns of President Joe Biden and former President Trump. These attacks, ongoing since May, have been directed at both current and former government officials, as well as campaign staffers.