AI

From Weak-to-Strong Alignment to Human-AI Co-Alignment to Sustainable Symbiotic Society

Authors:Yi Zeng, Feifei Zhao, Yuwei Wang, Enment Lu, Yaodong Yang, Lei Wang, Chao Liu, Yitao Liang, Dongcheng Zhao, Bing Han, Haibo Tong, Yao Liang, Dongqi Liang, Kang Sun, Boyuan Chen

View the PDF file from the paper entitled Refining the Exclusive Evaluation: From the weak compatibility to the strong to the joint alliance between the human-AA to the sustainable Takaful community, by Yi Zing and 15 other authors

PDF HTML (experimental) view

a summary:artificial intelligence Systems (AI) has become increasingly stronger and independent, and may exceed the levels of human intelligence, which are artificial services (ASI). While progressing from artificial intelligence to ASI, it may go beyond human control, violate human values, and even lead to irreversible catastrophic consequences in extreme cases. This leads to an urgent issue that must be addressed: superior evaluation, ensuring that artificial intelligence systems are more intelligent than humans, remains compatible with the intentions and mankind (compatible). Current developmental supervision and weak circular to strong to powerful can be unprecedented and insufficient when facing ASI. We must explore safer and more pluralistic frameworks and methods. In this paper, we redefine the superior evaluation as the common alliance between man AI towards a sustainable symbiotic society, and we highlight a framework that merges external censorship and fundamental proactive alignment. The superior evaluation must be based on external supervision on the final decision -making decision, and it is completed by the automatic and correction automatic evaluation, to achieve continuous compatibility with the advanced values ​​of humanity. The ultra -internal proactive evaluation is rooted in a deep understanding of the self, others, society, integrating self -awareness, self -reflection, sympathy for inferring human intentions automatically, distinguishing good from evil and thinking proactive in human welfare, and ultimately investigating through the dependent reaction. Incorporating the censorship that it moves externally with the pre-emptive proactive alignment enables sustainable symbiotic societies through the common alliances of human-AA, which paves the way for AGI safe and beneficial for good, for man, and for sympathy ecosystems.

The application date

From: Hibo Tong [view email]
[v1]

Thursday, April 24, 2025 09:53:49 UTC (337 KB)
[v2]

Friday, 25 April 2025 15:32:41 UTC (339 KB)

Don’t miss more hot News like this! Click here to discover the latest in AI news!

2025-04-28 04:00:00

Related Articles

Back to top button