ChatGPT inspires multidimensional research in linguistics

Source:Chinese Social Sciences Today 2023-05-31

Innovative breakthroughs in AI technology are driving dramatic changes in natural language processing. The emergence of AI language products, such as ChatGPT, is reshaping our understanding of AI technology, which ultimately relies on the development of basic language processing technologies. Therefore, it is important to not only monitor the future development of natural and artificial language, but also reflect on how current linguistic theories can serve technology and embrace the future.

The co-existence of natural and artificial language will be the future. ChatGPT is an AI language model that integrates AI language generation and language intelligence processing technologies. It is able to generate language through effective recognition, deep learning and full imitation of natural language information. It can also conduct human-computer interactive dialogue, machine translation and information summarization and extraction through learning, reasoning, judging and responding to natural language, thus transforming natural language into AI language.

The intelligent language generated by ChatGPT is essentially a “computer programming language with symbol processing and logical reasoning capabilities”; it is generated through extensive pre-training using natural language data. In most cases, intelligent language adheres to grammatical rules, has clear semantic orientation, and shows logical coherence. In this sense, natural language serves as knowledge and data sources for AI language, and provides important material for ChatGPT pre-training. On the other hand, AI language becomes a variant of natural language and an extension of language knowledge.

In a nut shell, natural language and AI language complement each other. Their deep integration and cooperation will become another breakthrough in the development of AI technology as well as a new research topic in linguistics and the tech industry.

The study of language theory is shifting from independent human exploration to human-machine cooperation. Before the emergence of AI, the mysteries of language were exclusively explored by humans. Several theoretical systems have been established with regards to language structure and the laws governing language evolution or language use, such as structural linguistics, grammar and pragmatics.

The advent of AI technology and ChatGPT in particular has made us aware of the limitations of human ability to study language. AI’s capability to intelligently, autonomously and automatically learn and process language from vast amounts of data exceeds that of humans, and it has enormous potential for deep learning and imitation of language. After all, the memory and processing abilities of any human linguist are limited. The full integration of human intelligence and AI computing power will become a major trend in theoretical linguistics.

The application of intelligent language must move from “Internet of Everything” to “Interaction of Everything.” Human society is entering the era of the Internet of Things and the Internet of Everything. The Internet has solved the problem of information exchange and communication between people, while the “Internet of Everything” is supposed to solve the problem of information and data exchange between people, things, as well as people and things.

However, with the current level of technological development and application, effective exchange of information and data between people and things has not been completely achieved. Devices often fail to fully understand and process natural language, and are thus unable to make decisions or execute instructions effectively. Fortunately, intelligent language technology is enabling the transition to “Interaction of Everything.” Moreover, such interaction is acquiring a more humane, affective and artistic nature.

AI language will change the ways language is learned and studied, while ushering humans into a world marked by “Interaction of Everything.” At the same time, AI will pose serious challenges to the governance of information dissemination in terms of ethics, copyrights, and security. Theoretical linguistics can contribute to the dialectical and rational thinking required to address these challenges.


Li Jia is from the School of International Journalism and Communication at Beijing Foreign Studies University.

Editor:Yu Hui

Copyright©2023 CSSN All Rights Reserved

Copyright©2023 CSSN All Rights Reserved