9 | 0 | 103 |
下载次数 | 被引频次 | 阅读次数 |
对最近一次图灵测试进行分析和解读,肯定人工智能已经通过了初级图灵测试,认为人类智能和机器智能双足鼎立的“双智社会”正在到来。以大模型为例,阐释人工智能取得的原理性突破以及存在的局限——实例性和弱共识性,并通过五个例子说明机器智能相对于人类智能的另类性,从而为双智社会的形成提供科学依据。探讨双智社会的基本问题——两种智能的相互关系和相对位置。讨论双智社会中的竞争与合作,提出竞争内涵的改变——竞争对手从同类变为另类,竞争目标从物质生存为主变为精神生存为主,竞争成效从更多收获变为更多贡献。给出人类智能在双智社会中的四大优势和保持主导性的三个条件。结论为双智社会的真正福利是倒逼人类向更高文明形态攀升。
Abstract:Analyzing and interpreting the latest experiment on Turing test, it is affirmed that artificial intelligence has passed the primary Turing test. This and other facts show that the dual intelligence society consisting of human and machinery intelligence is coming. Taking LLMs as an example, explain the fundamental breakthroughs and limitations of artificial intelligence: properties of instance and weak consensus, and illustrate through five examples the alternative nature of machinery intelligence compared to human intelligence, providing scientific basis for formation of the dual intelligence society. Exploring two fundamental issues of the dual intelligence society: the relation between and relative position of two kinds of intelligence. Discussing competition and cooperation in the dual intelligence society, revealing the main changes in the competition: competitors shift from being of the same kind to being alternative, the key competition goal shifts from material survival to spiritual survival, and the competition effectiveness shifts from more gains to more contributions. Identify four main advantages of human intelligence against machinery intelligence and three conditions for maintaining dominance of human intelligence in the dual intelligence society. The conclusion is that the true welfare of the dual intelligence society is to force human being to climb towards a higher form of civilization.
[1] Turing A M.Computing Machinery and Intelligence[J].Mind,1950,59(236):433-460.
[2] Turing A M,Intelligence Machinery[EB/OL].[1948](2025-08-01).https://turingarchive.kings.cam.ac.uk/.
[3] Jones C R,Bergen B K.Large Language Models Pass the Turing Test[EB/OL].[2025-05-31](2025-08-01).https://arXiv.org/abs/:2503.23674.
[4] 陈小平.大模型关联度预测的形式化和语义解释研究[J].智能系统学报,2023,18(4):894-900.
[5] 陈小平.大模型:从基础研究到治理挑战[J].中国人工智能学会通讯,2024(1):2-9.
[6] Arkoudas K.GPT-4 Can’t Reason[EB/OL].[2025-02-12](2025-08-01).https://arXiv.org/abs/:2308.03762.
[7] Yehudai G,Kaplan H,Ghandeharioun A,et al.When Can Transformers Count to n?[EB/OL].[2024-07-21](2025-08-01).https://arXiv.org/abs/:2407.15160.
[8] Huang K,Guo J,Li Z,et al.MATH-Perturb:Benchmarking LLMs’ Math Reasoning Abilities against Hard Perturbations[EB/OL].[2025-02-12](2025-08-01).https://arXiv.org/abs/:2502.06453v2.
[9] Ji J,Wang K,Qiu T,et al.Language Models Resist Alignment:Evidence from Data Compression[C]// Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1:Long Papers).Vienna:Association for Computational Linguistics,2025:23411-23432.
[10] 陈小平.智能革命将重塑怎样的世界[J].企业改革与发展,2024(11):18-22.
[11] 陈小平.人工智能三原则:探索机器智能与人类福祉的和谐共生之道[J].探索与争鸣,2025(3):7-10.
基本信息:
DOI:10.19484/j.cnki.1000-8934.2025.09.002
中图分类号:TP18
引用信息:
[1]陈小平.双智社会的颠覆性变革前瞻[J].自然辩证法研究,2025,41(09):26-33+50.DOI:10.19484/j.cnki.1000-8934.2025.09.002.
基金信息: