Brain-Machine Intelligence and the Problem of Representation
04-01-2026

Social Sciences in China (Chinese Edition)

No. 2, 2026

 

Brain-Machine Intelligence and the Problem of Representation

(Abstract)

 

Li Zhongwei

 

Brain-machine intelligence (BMI) integrates biological intelligence and artificial intelligence through brain-computer interfaces. Representation is a defining feature of natural minds and a core concept in BMI research. Human representations directly refer to environments, objects, and their affordances, and are often conscious. By contrast, current representations in forms of BMI largely lack world semantics and conscious experience. Recent technological advances and philosophical reflections suggest the need to reconceptualize representational systems in BMI. Specifically, representational vehicles should be understood as internal mechanisms integrating neural signals, chips, algorithms, and the states of mechanical components, while representational targets should consist of environments, objects, and their affordances. Such a reconceptualization would enable BMI systems to interact more effectively with their environments and objects, or allow emotions to be better attuned to their environments and objects. At the same time, appropriate forms of conscious experience should be realized through approaches such as closed-loop brain-computer interfaces. On this basis of the mechanisms underlying representation and consciousness, natural intelligence and artificial intelligence can be more deeply integrated, prompting renewed philosophical reflection on subjectivity, autonomy, and identity, as well as sustained attention to the ethical implications of BMI.