The danger of AI will come from its self-consciousness rather than its capacity. AI would be a challenge to mankind if it develops a capacity for reflection on the system of its own, so that it would know how to remake its system with new rules, especially when it could invent its own language for all purposes, equal in capacity to human natural language, it could do anything it would. And AI would be most dangerous if it would be able to learn human desires, emotions and values, because all hostilities are based upon human desires and values.

Turing machine; Godel; Humanization

Citation: Zhao, T. 'How Could AI Develop Its Self-consciousness?' [J]. Journal of Dialectics of Nature, 2019, 41(1): 1-8.