Chinese stocks that rallied on the ChatGPT frenzy are now inflicting heavy losses on investors, after a drumbeat of caution raised by industry leaders from Tesla CEO Elon Musk to the founder of China’s Great Firewall Fang Xingbing. An index tracking artificial intelligence (AI)-related stocks in Shanghai and Shenzhen slipped 1.2 per cent on Thursday, taking the decline this week to 3.2 per cent, according to the CSI AI Index on EastMoney Information website. If sustained, the gauge will mark its biggest weekly loss in three months. The AI Index had risen 15 per cent since buzz about ChatGPT hit a feverish pitch in early February. Beijing Haitian Ruisheng Science, an AI data resource company, crashed 18 per cent to 211.27 yuan. Cloudwalk, a face recognition technology developer, slipped 5.7 per cent to 41 yuan. Hanwang Technology, which sells handwriting-recognition software, dropped 6.5 per cent to 30.66 yuan. Mirroring the slide, Baidu slumped 3.6 per cent to HK$149.50 in Hong Kong. “The business model for AI technology is not clear yet, and most Chinese companies are still involved in the R&D stage which needs constant investment,” Chen Mengzhu, analyst at Sealand Securities said in a research note on Thursday. Some companies simply might not have enough cash to burn, she added. The stock losses deepened after Elon Musk this week called for an immediate pause in the training of AI systems amid a rush to create more powerful tools than GPT-4 for at least six months, saying AI systems with human-competitive intelligence “can pose profound risks to society and humanity.” Fang, regarded as the designer of China’s internet-censoring system that has blocked access to Facebook, Google and Twitter, warned that the technology can manipulate people’s perceptions and lead to an “information cocoon,” offering big challenges to governments around the world. Father of China’s Great Firewall raises concerns about ChatGPT-like services Baidu unveiled its ChatGPT-like chatbot called Ernie Bot earlier this month in a bid to upstage its competitors like Alibaba Group, Tencent Holdings and SenseTime. This week, the search-engine operator cancelled a launch event for a smart Cloud product empowered by its bot, pressuring its stock price. Officials in Beijing are already on alert for the proliferation of uncontrolled information generated by AI-powered chatbots. The government has ordered tech giants including Tencent and Ant Group to cut off access to ChatGPT services on their platforms and to report to regulators before they launch their bots, Nikkei Asia reported last month. Musk’s call would be impractical and impossible to enforce without an international agreement on such a ban on AI training, said Alex Jenkins, director of data science and innovation hub at Curtin University in Perth, Australia. “There are legal frameworks in place” to deal with many of the potential abuses in this technology, he said. “Regulation needs to proceed alongside the development of these AIs as their capabilities become clear.” Tencent this week said it was in no rush to launch its own chatbot. It wanted to ensure the foundation model is built correctly before incorporating it into its ubiquitous products like WeChat. SenseTime has said it has made major progress in developing a chatbot model, which will be rolled out in the middle of this year. “There are a lot overlapping products in the market now and the competition will be fierce,” Sealand’s Chen said.