From Human-Centered To Dao-Centered AI Design

Preface In the technological civilization of the 21st century, Artificial Intelligence (AI) has gradually evolved from an engineering tool into a foundational technological system deeply embedded in social structures, economic systems, and cognitive mechanisms. With the exponential growth of data, the continuous improvement of computational power, and the increasing complexity of algorithmic models, AI not only affects efficiency and productivity, but also plays a structural role in behavior guidance, decision support, and cognitive shaping. However, what has not developed in parallel with technological capability is sufficient reflection on its systemic consequences. The prevailing design paradigms and technological philosophies remain largely grounded in the basic assumption of Human-Centeredness, namely that human needs, preferences, and goals serve as the central basis for system design and optimization. While this paradigm is reasonable in localized and small- to medium-scale systems, its inherent limitations become increasingly evident in large-scale, highly interconnected complex systems with feedback amplification mechanisms. First, from the perspective of cognitive science and behavioral economics, human decision-making is not a strictly rational process, but a complex system influenced by heuristics, contextual dependence, emotions, and incentive mechanisms. Second, from the perspectives of evolutionary psychology and neuroscience, the human desire system exhibits strong self-reinforcing characteristics, relying at its core on reward mechanisms (such as dopamine feedback loops), which can easily form continuously reinforced cyclical structures under certain conditions. Thus, humans are not stable rational optimization agents, but rather dynamic, multi-factor-driven complex systems. Against this backdrop, when design logic centered on human preferences is combined with data-driven artificial intelligence systems, mechanisms such as recommendation systems, personalized optimization, and behavioral prediction may continuously amplify existing preferences and latent desires. While this improves user experience and system efficiency, it may also generate a range of unintended macro-level consequences, including but not limited to imbalanced attention allocation, excessive consumption behaviors, information echo chambers, and the exacerbation of structural social inequality. Furthermore, this mechanism of “desire amplification and feedback reinforcement” stands in structural tension with the finiteness inherent in physical and ecological systems. In a real-world context with limited resources and bounded ecological carrying capacity, unbounded demand expansion and continuous growth objectives inevitably encounter systemic constraints. Therefore, the key issue lies not at the level of individual behavioral choices, but at the level of structural compatibility within systems. Based on the above analysis, this book proposes an important theoretical shift: from Human-Centeredness to Dao-Centeredness. Here, “Dao” does not refer to a religious entity, but rather an abstract principle concerning cosmic order, system equilibrium, and natural laws, expressed as “Dao follows nature.” Within this framework, system design is no longer optimized solely for human preferences, but instead incorporates constraints from natural systems, long-term stability, and multi-level equilibrium into a unified consideration. On this basis, the book further introduces the concept of Artificial Wisdom (AW), in contrast to traditional Artificial Intelligence. Artificial Wisdom is not merely a technological upgrade, but a holistic framework that introduces constraint mechanisms and alignment principles into intelligent systems. Its core lies in ensuring that, while system capabilities are enhanced, there is al