AI models require regulation in domains such as cybersecurity and virology, particularly in a Chinese context