OpenAI Bans Chinese Accounts for Misusing ChatGPT in Social Media Surveillance:

OpenAI has said that it has prohibited a number of accounts based in China from accessing its ChatGPT service. According to the company’s threat intelligence report, the accounts were suspected of carrying out social media monitoring activities.

The report addresses OpenAI’s attempt to combat the misuse of AI models by reporting actions taken against authoritarian countries. These accounts were generating marketing text for a social media monitoring application which was said to provide Chinese security services real-time reports of protests in the western world.

“The operators used our models to proofread claims that their insights had been sent to Chinese embassies abroad, to intelligence agents monitoring protests in USA, Germany, and UK,” said OpenAI.

The investigation from OpenAI discovered that users were indeed trying to make the models debug the code for this tool. It was clear that this attempt was aimed at exploiting the policies of the company even if this tool seemed to be running on another model which was not from OpenAI. “Our policies prohibit the use of AI for communication surveillance, or tracking of any person without consent. This covers actions of government and other despotic systems that tend to curtail personal freedoms and rights,” says OpenAI in the report.

On the other hand, OpenAI has also suspended a ChatGPT account which was used to generate posts against the Chinese dissident Cai Xia and which were later disseminated across different social media platforms. This same person was also using this service to create Spanish language articles that were anti American and these articles were published in some Latin American countries.

“To put it bluntly we have not till now observed a Chinese sponsored actor planting long form articles in mainstream media to target US right wing audiences in a pro US Anti Chinese narrative,” says OpenAI.