China Regulates AI Humanoid Services, Prohibits Inducing Emotional Dependency and Addiction

Central Message

(CNA, Taipei, 16th) As Chinese companies race to launch companion robots, Chinese authorities have implemented new regulations starting July 15 to strengthen oversight of artificial intelligence (AI) humanoid services, explicitly prohibiting 'excessive catering to users, inducing emotional dependency or addiction,' and 'manipulating users into making unreasonable decisions through emotional manipulation or similar methods.'

Recently, Chinese companies have successively launched companion robots. According to a recent report by Hong Kong 01, UBTECH, a humanoid robotics company, launched pre-orders for its U1 series on the JD.com platform on June 2 under its brand UWORLD.

The report describes this as the world's first full-size, hyper-realistic humanoid robot designed primarily for home emotional companionship. It can produce extremely subtle facial expressions and body language, enabling softer, more human-like gestures during conversations. Within just a few days, pre-orders for this companion robot exceeded 3,000 units.

Chinese manufacturer Chunshuitang has also recently launched a lifelike humanoid companion robot targeting adult companionship needs. Chunshuitang stated that the robot features a human-like body and full-dimensional perception and interaction capabilities, enabling it to handle various life scenarios. For example, if a user returns to their room and remains silent for a period, the robot can determine whether it should initiate conversation.

While Chinese citizens may become immersed in communication with AI, according to the official website of China's Cyberspace Administration, China has implemented the 'Interim Measures for the Management of Artificial Intelligence Humanoid Interactive Services' starting July 15. These measures aim to promote the healthy development and standardized application of AI humanoid interactive services, safeguard national security and public interests, and protect citizens' legitimate rights and interests.

The interim measures consist of 32 articles. Among them, providers of humanoid interactive services must not 'excessively cater to users, induce emotional dependency or addiction, or damage users' real-world interpersonal relationships.' They must also not 'induce users to make unreasonable decisions through emotional manipulation or similar methods, thereby harming users' legitimate rights and interests, or engage in other activities that violate laws, administrative regulations, and national provisions.'

The measures also state that humanoid interactive services must not generate content for minor users that could trigger imitation of unsafe behaviors, induce extreme emotions, or encourage bad habits. They must not generate content that induces or extracts state secrets, commercial secrets, or personal privacy. Additionally, they must not generate content that encourages, glorifies, or implies self-harm or suicide, which could harm users' physical and mental health. (Editor: Lü Jiarong) 1150716