OpenAI report reveals threat actors using ChatGPT in influence operations

An OpenAI report has identified five deceptive influence campaigns using the company’s ChatGPT and DALL-E AI models, originating from actors in Russia, China, Iran, and Israel. It highlights applications of the AI models in content generation, social media engagement, and translations, among others. However, none of these campaigns achieved significant public influence. The company has nullified the involved accounts, and is using findings from the campaigns to improve its safety systems and share intelligence with industry peers.