Evaluating AI-Generated Language as Models for Strategic Competence in English Language Teaching
Saved in:
| Title: | Evaluating AI-Generated Language as Models for Strategic Competence in English Language Teaching |
|---|---|
| Language: | English |
| Authors: | Phuong-Anh Nguyen |
| Source: | IAFOR Journal of Education. 2024 12(3):325-349. |
| Availability: | International Academic Forum. Sakae 1-16-26 - 201 Naka Ward, Nagoya Aichi, Japan 460-0008. Tel: +81-50-5806-3184; Web site: http://iafor.org |
| Peer Reviewed: | Y |
| Page Count: | 25 |
| Publication Date: | 2024 |
| Document Type: | Journal Articles Reports - Research |
| Descriptors: | Artificial Intelligence, Computer Software, Metacognition, Technology Uses in Education, Teaching Methods, Communication Strategies, Transcripts (Written Records), Taxonomy, Cues, Interaction, Instructional Materials, Computational Linguistics, English (Second Language), Dialogs (Language), Task Analysis |
| ISSN: | 2187-0594 |
| Abstract: | Strategic competence, the ability to use communication strategies (CS) to overcome challenges and enhance communication effectiveness, is crucial in language learning. However, the coverage of these strategies as well as target models to teach them remain scarce in current instructional materials. This paper represents the first attempt to examine the application of ChatGPT in providing target models of CSs and facilitate L2 learners' strategic competence. ChatGPT-4 was used to generate transcripts of monologues and dialogues around a description task following two types of prompts: with and without a taxonomy of communication strategies (structured and unstructured prompts). Preliminary findings suggest Chat-GPT's considerable potential in modeling communication strategies. Across the two prompting conditions, the chatbot was able to present a wide range of CSs, including achievement, self-monitoring, timegaining, and interactive strategies. The highest CS content was found in the structured-prompt dialogue which utilized 9 out of 10 CS sub-types, a more diverse range than typically covered in textbooks, with approximation, circumlocution, and time gaining being most frequently used. In terms of linguistic presentation, the AI-generated transcripts demonstrated appropriate use of CSs, though their linguistic realizations were limited in range. The article concludes with implications for leveraging Chat-GPT to contextualize communication strategies, considerations for prompt engineering, strategy training to proficiency levels, and AI-teacher collaboration. |
| Abstractor: | As Provided |
| Entry Date: | 2024 |
| Accession Number: | EJ1453502 |
| Database: | ERIC |
| Abstract: | Strategic competence, the ability to use communication strategies (CS) to overcome challenges and enhance communication effectiveness, is crucial in language learning. However, the coverage of these strategies as well as target models to teach them remain scarce in current instructional materials. This paper represents the first attempt to examine the application of ChatGPT in providing target models of CSs and facilitate L2 learners' strategic competence. ChatGPT-4 was used to generate transcripts of monologues and dialogues around a description task following two types of prompts: with and without a taxonomy of communication strategies (structured and unstructured prompts). Preliminary findings suggest Chat-GPT's considerable potential in modeling communication strategies. Across the two prompting conditions, the chatbot was able to present a wide range of CSs, including achievement, self-monitoring, timegaining, and interactive strategies. The highest CS content was found in the structured-prompt dialogue which utilized 9 out of 10 CS sub-types, a more diverse range than typically covered in textbooks, with approximation, circumlocution, and time gaining being most frequently used. In terms of linguistic presentation, the AI-generated transcripts demonstrated appropriate use of CSs, though their linguistic realizations were limited in range. The article concludes with implications for leveraging Chat-GPT to contextualize communication strategies, considerations for prompt engineering, strategy training to proficiency levels, and AI-teacher collaboration. |
|---|---|
| ISSN: | 2187-0594 |
