In May, OpenAI released research revealing operations linked to Russia, China, Iran, and Israel had been using its tools to create and spread disinformation.

Source: Political deepfakes are the most popular way to misuse AI | Ars Technica