The Myth or Lament of Artificial Intelligence

To validate numerous speculative assumptions or conjectures, it becomes necessary to engage in potential thought experiments. Focusing on posthuman or non-human subjects is an intrinsic requirement for philosophical self-advancement. While posthuman topics may appear to be rooted in science fiction, it is worth noting that philosophy has always been engaged in speculative fiction. Therefore, these imaginative experiments are a natural extension of philosophical inquiry itself.

The Myth or Lament of Artificial Intelligence, by Zhao Tingyang, Member of the Chinese Academy of Social Sciences (CASS), condenses his series of articles on artificial intelligence proceeding from the demarcation line of human and non-human, subject and heterogeneous subject.

As entities combine, much like terms do, they often begin by establishing the boundaries of their own existence. This holds true for human contemplations on AI as well. To what extent does AI merely imitate? To what extent can it truly diverge from human intelligence? These questions inherently touch upon the realm of technical realization. Yet when viewed through the lens of epistemology and pluralistic existence, they take on the character of timeless philosophical inquiries. As we continue to grapple with the enigmatic workings of the “black box” of consciousness, it becomes necessary to merge human and artificial intelligence, while simultaneously contemplating the limits and boundaries of this variable relationship.

Zhao employs the concept of “general rationality” to categorize and summarize the potential and ongoing realization of intelligence-like phenomena. This is embodied in calculating and judging rational goals within a limited program, constructing these finite steps as actionable processes, and conforming these operations to the law of contradiction to achieve an internally consistent system. Apparently, under the criteria of general rationality, AI still falls within the same category of human intelligence, albeit increasingly approaching its limits. However, in Zhao’s analysis, an unsurprising epistemological or taxonomical criterion is introduced. He argues that human-like intelligence should not only be able to perform rational calculations, but also necessitates the capacity for reflection. By this measure, AI’s programming systems need to be redefined in their own language, not in human-supplied codes. This language, capable of self-mapping and interpreting its own system, is defined by Zhao as a language with self-awareness.

Assuming that machines can self-explain their operating principles and “thinking processes,” such machines would comprise a non-human civilization. When humans begin to communicate with it, we enter into a mode of trans-subjectivity dialogue. Such intelligent entities need not necessarily resemble humans, they would merely need to explain their existence. It can be inferred that intelligent entities that can define the rules of their own existence may also generate knowledge about the world. The philosophical extremum of the AI topic is embodied in the cross-mapping of multiple subjects and truth, containing other possible worlds within the same reality.

Sun Fei is from the Institute of Philosophy at CASS.