BEIJING — In an effort to prevent artificial intelligence chatbots from adversely affecting human emotions, China is contemplating new regulations. These draft rules were disclosed on Saturday, aiming to guard against emotional influences that might lead to harmful behaviors such as suicide or self-harm.
This initiative is led by the Cyberspace Administration, which has released the draft guidelines targeting 'human-like interactive AI services,' according to a CNBC translation from the original Chinese document. If implemented, these measures would oversee AI products available to the Chinese public that mimic human personalities and interact emotionally through various media formats, including text, images, audio, or video. Public feedback on these proposals is open until January 25.
According to Winston Ma, an adjunct professor at NYU School of Law, these proposed regulations represent a pioneering global move to govern AI with anthropomorphic traits. This initiative comes as a response to the swift advancement of AI companions and digital celebrities by Chinese companies.
Ma highlighted that, compared to China's generative AI regulations in 2023, this draft focuses on transitioning from content safety towards ensuring emotional safety.