Specialised Discourse With Agents? Beyond ChatGPT in Academic English

Abstract

This chapter follows the current discussion of Artificial Intelligence, agents and Large Language Models (LLMs), especially ChatGPT, in academic writing and gives examples of how LLMs and other agents could be integrated into specialised (English) discourse now and in future. It briefly explains the linguistic components of agents and relevant LLMs as well as the idiomatic and coherent generation of output. Agents can be integrated into the whole spectrum of specialised, scientific research, from initial brainstorming to final editing, as many examples show. This chapter focuses on the understanding of LLMs for specialised discourse and asks crucial questions, such as whether LLMs write better than humans and whether AI-generated work can be detected automatically. This contribution looks beyond ChatGPT in two perspectives: it demonstrates the wide range of LLMs that have become available after ChatGPT appeared in late 2022 and discusses future possibilities, such as integrating a spoken component into LLMs such as a conversational agent that speaks non-standard English (e.g. Czech or Italian English). In any case, although agents can be used efficiently for routine language work, particularly by non-native specialists, they are unlikely to replace well-trained academic teachers and researchers, but AI-assisted research could replace non-AI-assisted research.

Ähnliche Arbeiten