Generative AI, Pragmatics, and Authenticity in Second Language Learning

Robert Godwin-Jones`

2024-10-18

Abstract:There are obvious benefits to integrating generative AI (artificial intelligence) into language learning and teaching. Those include using AI as a language tutor, creating learning materials, or assessing learner output. However, due to how AI systems under-stand human language, based on a mathematical model using statistical probability, they lack the lived experience to be able to use language with the same social aware-ness as humans. Additionally, there are built-in linguistic and cultural biases based on their training data which is mostly in English and predominantly from Western sources. Those facts limit AI suitability for some language learning interactions. Stud-ies have clearly shown that systems such as ChatGPT often do not produce language that is pragmatically appropriate. The lack of linguistic and cultural authenticity has important implications for how AI is integrated into second language acquisition as well as in instruction targeting development of intercultural communication compe-tence.

Computation and Language,Artificial Intelligence,Human-Computer Interaction

What problem does this paper attempt to address?

The paper attempts to address the application and limitations of generative artificial intelligence (AI) in second language learning. Specifically, the paper explores the following points:

1. **Language capabilities of generative AI**: Although generative AI (such as ChatGPT) performs well in language generation, capable of producing grammatically correct and structurally sound text, these systems lack real human social experience and cultural background, leading to significant shortcomings in handling the social and cultural dimensions of language.

2. **Authenticity and cultural sensitivity of language**: The output of generative AI often lacks authenticity and cultural sensitivity. Since the training data mainly comes from Western English sources, these systems may exhibit cultural bias and inaccuracies when generating content in other languages.

3. **Pragmatic issues of language**: Generative AI performs poorly in handling the pragmatics of language. Pragmatics focuses on the appropriate use of language in specific social and cultural contexts, and generative AI, due to its lack of human social experience, struggles to demonstrate appropriate pragmatic abilities in actual conversations.

4. **Impact on second language learning**: The paper discusses the potential role and limitations of generative AI in second language learning. Although AI can serve as an auxiliary tool for language learning, providing practice and feedback, its limitations at advanced levels of language learning may affect learning outcomes, especially in developing cross-cultural communication skills.

In summary, the paper aims to explore the potential and challenges of applying generative AI in second language learning, particularly its limitations in terms of language authenticity, cultural sensitivity, and pragmatics.