AI-assisted Term Management on Professional Learning Platforms: Methodogical Keys for Meaning Description Adjustment

Formulating optimal meaning descriptions of terms in professional, terminographic reference tools is a much debated and complex issue, see for instance Béjoint and Thoiron (2000) for an overview. Even more so, on a professional learning platform, meaning descriptions included in a terminographic reference module should be adjusted according to the didactics and pragmatics of the organization in line with the specific learning needs of the users, while at the same time being truly supported by viable business strategies and models, Simonsen (2023a).

With the advent of powerful generative AI tools, see for instance ChatGPT (2024), and the realization of the potential of AI-assisted term meaning description tasks, see for instance de Schryver & Joffe (2023), there seems to be a need for additional theoretical and methodological reflections on AI-assisted term management for use on professional learning platforms. How can AI help terminographers generate and manage adjusted meaning descriptions? Which guidelines should be used for prompting the formulation of rule-based instructions and thus get the best learning output of the technology?

The purpose of this paper is to discuss and demonstrate how different methodological approaches can be favourably used and combined to support and facilitate AI-assisted meaning description generation in a professional learning environment. We will explain how different theoretical keys unlock different options for meaning description and meaning adjustment for the benefit of the various users of the digital learning environment and for the organisation.

Our paper draws on insights from a critical literature review of relevant theoretical contributions from different research fields. The first theoretical key comes from the semantically driven approach as suggested by Wiegand (1994). The second key comes from the encyclopaedic approach also coined by Wiegand (1994). The third key comes from the functional approach described by Leroyer (2018) for the purpose of text reception situations, and also suggested by Agerbo & Bergenholtz (2017), who argue for the use of a similar functional approach of operative situations and cognitive situations, respectively. The fourth key is the pragmaterminological approach, de Vecchi (2018, 2020), who makes the case for corporate terminography using four elements including forms, users, validity over time and actions. The fifth key is the LMCP approach, Leroyer & Simonsen (2020), which adds the necessary layers of meaning construction in organisations. The final and sixth key is the lexicographic prompting approach for generative AI tools, Simonsen (2023b).

In addition, the paper also draws on empirical insights from a qualitative case study of meaning descriptions generated in a private learning organization. A sample of insurance term descriptions are analysed by means of a user survey involving both students and teachers. The purpose of the test is to assess the learning efficacy of combinations of different theoretical keys.

The analysis establishes that different combinations can be successfully used in the strategic as well as in the practical day-to-day work of term management in the case study company. The analysis also led to the design of the meaning description adjustment hexagon, which outlines the six different approaches and understandings of meaning construction and meaning adjustment of terms in a professional, organizational context. It is argued that the meaning hexagon can be effectively used to manage meaning adjustment and is a highly flexible tool for the application of generative AI in the development of curated terminographic resources in accordance with the organizational mission and the students’ learning styles and objectives.