OpenAI Blocks ChatGPT Accounts Allegedly Controlled By The Chinese Government, Asks To Create Spy Tools For Social Media

JAKARTA - OpenAI announced that it had blocked a number of ChatGPT accounts suspected of having links to Chinese government entities, after users of these accounts asked for the chatbot to compile proposals related to monitoring conversations on social media.

In its latest public threat report, the San Francisco-based company said that some individuals had used ChatGPT to request a draft social media listening as well as other digital surveillance concepts that violated OpenAI's national security policy.

The move highlights new concerns about the abuse of generative artificial intelligence, amid increasingly intense competition between the United States and China in developing and regulating AI technology.

In addition to accounts connected to China, OpenAI also closed a number of Mandarin-language accounts known to use ChatGPT to assist phishing campaigns and the manufacture of malware, and asked the AI model to examine additional automation that can be done through a Chinese AI system called DeepSeek.

The Chinese Embassy in the United States has yet to comment on this report.

Not only that, but OpenAI also disabled accounts suspected of being related to Russian-language criminal groups, who used ChatGPT to help develop certain types of malware.

According to OpenAI, since it began publishing public threat reports in February 2024, it has thwarted and reported more than 40 malicious networks. The Microsoft-backed company insists that their model does not show any new tactics or provides new offensive capabilities to threat actors.

"We find no evidence of any new tactics or indications that our model provides unprecedented attack capabilities," OpenAI wrote in its report.

OpenAI, which now has more than 800 million ChatGPT users per week, has just become the world's most valuable startup with a valuation of 500 billion US dollars after completing the sale of secondary shares last week.