China plans to test the security of all generative neural network-based services before allowing them to operate in the country. This raises some concerns about the future of chatbots like Open AI’s ChatGPT, developed by local IT giants like Baidu.
Service providers must ensure that AI algorithms provide accurate information and respect copyrights and do not compromise user security. Developers of AI algorithms must also clearly label the content generated by the neural network. So says the draft by the Cyberspace Administration of China, which has been submitted to the general public for public feedback.
The Chinese Cyberspace Administration’s demands come as part of larger efforts by the authorities to regulate the generative neural network market, which has exploded since the launch of the ChatGPT algorithm last November. Chinese IT companies like Alibaba, SenseTime and Baidu aim to build a robust neural network platform for the world’s largest internet marketplace.
Against the background of the regulator’s statement, Alibaba’s share price fell significantly, while SenseTime’s securities almost did not lose value. Earlier this week, SenseTime announced its own AI bot, SenseChat, positioned as a competitor to ChatGPT. Baidu started testing its Ernie chatbot about a month ago. Currently, Baidu is considered the leader in China’s generative neural network market.