OpenAI bans Russian, Iranian accounts over influence efforts : NPR

Key Points:

• OpenAI banned Russian and Iranian accounts using its AI tools, including ChatGPT, to conduct influence operations that spread false narratives in multiple countries, with some misinformation reaching established media and prompting political responses.
• Russian operations targeted Latin America, aiming to undermine support for Ukraine and influence local politics by impersonating authorities and staging events, such as a controversial Ukraine-related school activity in Peru that sparked reactions in Poland.
• The Russian campaigns primarily used ChatGPT to generate internal reports for superiors, while only a minority of their content was AI-generated, and some efforts involved running fake think tanks and taking credit for unrelated events.
• Iranian operatives used AI to produce and publish nearly 100 articles under seven fake authors in Persian and English, targeting U.S.-Iran relations and American politics, with content appearing on outlets like Daily Kos and Middle East Monitor.
• Experts noted that AI helped these influence operations create more authentic-seeming content, making it easier to infiltrate legitimate media outlets and complicate detection compared to previous campaigns.