OpenAI, a leading artificial intelligence company, has revealed that it has shut down five covert influence operations in the past three months. These networks, located in countries such as Russia, China, Iran, and Israel, were using OpenAI’s products to manipulate public opinion and shape political outcomes while concealing their true identities.
In a recent report, OpenAI highlighted how these influence networks utilized AI tools to generate text and images at a larger scale and with fewer errors than human-generated content. Despite their efforts, the company stated that these campaigns ultimately failed to significantly increase their reach.
Ben Nimmo, principal investigator on OpenAI’s Intelligence and Investigations team, emphasized the importance of addressing the potential risks associated with AI-powered influence operations. He stated, “With this report, we really want to start filling in some of the blanks.”
The identified covert operations, including groups like “Doppelganger” and “Spamouflage,” used OpenAI’s AI models to automate posting on platforms like Telegram and generate comments to sway public opinion. OpenAI also noted that some of the AI-generated content contained common error messages, allowing them to identify and defend against such operations.
While these influence networks may not have gained widespread traction, OpenAI stressed the need for continued vigilance. Nimmo warned, “History shows that influence operations which spent years failing to get anywhere can suddenly break out if nobody’s looking for them.”
OpenAI’s efforts to combat covert influence operations align with similar disclosures made by companies like Meta Platforms Inc. The company plans to share more reports in the future and collaborate with industry peers to address the growing threat of AI-powered manipulation.