At his best, man is the noblest of all animals, separated from law and justice he is the worst.
Aristotle
The current trajectory of technology, and the future of society itself, are being profoundly shaped by Artificial Intelligence and other forms of anthropomorphic technologies (technologies that possess human traits and thought patterns, without actually being human). The power of this relatively recent technology, combined with the speed at which it is evolving, could become a genuine recipe for disaster if left unchecked.
If we are honest, we have already seen the impact AI can have in promoting illegal or unhealthy activities. And with each passing day, the capabilities of anthropomorphic technologies continue to grow. To maintain any form of control in the future, clearer guidelines must be established, particularly in regions such as the European Union and the United States.
China has recently launched a public consultation on a new law aimed precisely at regulating these technologies. The proposal is titled “Interim Measures for the Administration of Humanized Interactive Services Based on AI”, and its purpose is clearly reflected in its name.
The draft legislation consists of 32 articles, all of which set stricter and more clearly defined expectations for both AI software and its developers, with the goal of protecting the public interest. Despite the significance of this announcement, it has received limited attention, especially in regions such as the EU and the US, where existing guidelines remain vague.
The articles range from technical requirements for software to safety and health standards for users. Let’s look at some of the key highlights.
Article 2 focuses on raising awareness of what AI actually is. It defines artificial intelligence as software that contains human traits, can think like a human, and can encourage ideas similar to those of a human being. The first step toward regulating AI is understanding its nature, and this article establishes that foundation.
Article 3 builds on this concept, emphasizing that despite the advanced state of AI development, the Chinese government still seeks to promote creativity in anthropomorphic technologies, but within a safe and controlled environment.
Article 7 shifts the focus directly to the software itself. It outlines the standards that any anthropomorphic technology must meet in order to operate legally. These restrictions include prohibitions on:
It also establishes limits on how software may influence human behavior, banning the use of algorithms to manipulate users or encourage dangerous behaviors, including self-harm or violence toward others.
In essence, this article treats anthropomorphic software as if it were human. If it breaks the law by promoting violence or causing unrest, it is considered a legal violation with consequences equivalent to those faced by individuals.
Articles 8 through 18 address responsibility for the actions of the software. Articles 8 and 9 state that full responsibility for safety and security lies with the provider throughout the entire lifecycle of the software. Providers must also have the technical capacity to fully manage their systems.
Articles 11 and 12 focus on the protection of minors and elderly users. All providers are required to include a “minor mode” and obtain consent from a legal guardian. For elderly users, the regulation encourages the inclusion of emergency contact information in case an issue arises during use.
Articles 16 through 18 establish transparency requirements. Providers must clearly inform users that they are interacting with an artificial intelligence system, regardless of the software’s appearance or behavior. AI must not replace human social interaction, and its use should be limited.
Additionally, the software must be capable of detecting problematic usage. For example, after two consecutive hours of interaction, it must notify the user that they have been active for an extended period and remind them that they are interacting with AI, not a human. Users must also be able to exit the interaction at any time, in a simple and fully accessible manner.
Finally, Article 29 outlines the consequences of violating these guidelines. In practice, anthropomorphic systems are treated as if they were individuals, although liability ultimately falls on the developing company. Repeated violations may result in penalties, including the suspension of the software.
Given that neither the EU nor the United States currently has strict regulations governing AI use, it may be time to learn from China’s approach. While China has openly stated its ambition to lead the AI race and operates within a system where such rules may be easier to enforce, establishing clear boundaries is a necessary step for the future of technology.
Anthropomorphic technologies grow smarter and more sophisticated every day. Controlling them before it is too late is essential. A similar regulatory framework would not only help mitigate risks, but also deepen our understanding of these technologies. If we want an efficient and sustainable future, we must know how to use artificial intelligence and, more importantly, how to control its use.
