An increasing number of Chinese groups are using the ChatGPT chatbot to conduct covert operations, the US company OpenAI, the producer of the software, said in a report published on Thursday.
The report found that the scope and range of tactics used by these groups had expanded, although the operations identified were small and targeted limited audiences.
OpenAI has banned ChatGPT accounts from its platform that generate social media posts on political and geopolitical topics relevant to China , including criticism of video games focused on Taiwan, false accusations against a Pakistani activist, and content related to the closure of the US government aid agency USAID.
Some of the content criticized U.S. President Donald Trump's tariff increases , with posts on the X platform including: “Tariffs are making imported goods outrageously expensive, yet the government is squandering foreign aid money. So who's going to eat?”
In another operation, China-linked entities supported various stages of their cyber operations, including searching open sources, modifying scripts and configuring bug fixing systems.
The third case OpenAI found involved an influence operation to create polarizing social media content , such as supporting both sides of the US political discourse, using AI-generated profile photos.
OpenAI regularly publishes reports on malicious activity detected on its platform, such as the creation of malware or fake content on websites and social media. (PAP)
mw/ akl/
