OpenAI has revealed in a new security report the removal of several ChatGPT accounts used by China-based networks. One of them was used to generate Spanish-language articles denigrating the United States. They were published in mainstream media in Latin America, and signed by a Chinese company.
AI under influence
Other malicious actors, this time linked to North Korea, have exploited ChatGPT's AI to generate resumes and online profiles of fictitious candidates, in order to fraudulently land jobs in Western companies. Another network, apparently connected to a financial fraud operation in Cambodia, used the technology to translate and generate comments on social media sites like X (formerly Twitter) and Facebook.
OpenAI also banned a set of accounts that used its AI models to monitor online conversations on social media, in order to identify discussions on sensitive topics related to China, including calls for human rights protests. The information collected in several countries (the United States, the United Kingdom and Germany) was clearly intended for Chinese authorities.
ChatGPT is reportedly unavailable in China, after OpenAI blocked its services last summer. But Microsoft continues to offer the model through Azure, its cloud platform. U.S. authorities have already says fear China’s use of AI to repress the population and spread disinformation.
Source: Reuters
0 Comments