Empowerment or Illusion? Generative AI and the Promise of Inclusion for EFL Learners with Learning Difficulties: Teachers’ Perspectives

Abstract

English as a Foreign Language (EFL) instruction is making use of generative artificial intelligence (AI) tools such as students with learning difficulties including dyslexia, attention-deficit/hyperactivity disorder (ADHD), or processing difficulties. The effectiveness of these tools as true empowerment as opposed to a superficial appearance of support has not yet been tested empirically from the teachers' point of view. The present research was conducted with 91 EFL/ESL teachers with varying levels of education and from diverse countries who taught at least one student with a learning difficulty in the last two years. A convergent parallel mixed-methods approach was used, where teachers completed a validated Empowerment Scale (with excellent internal consistency, α = .899) and Skepticism/Illusion Scale (with excellent internal consistency, α = .955), and open-ended questions were analyzed using thematic analysis. Empowerment and Skepticism scores were modestly, negatively correlated, r (89) = -.258, p = .014; however, there was no correlation between the training and the reduction of Skepticism. Teachers who participated in prior training reported significantly higher levels of Empowerment than those who did not, t (89) = 2.48, p = .015, d = 0.65. Thematic analysis showed that nearly all teachers did not label Generative AI as simply enabling or posturing, but rather as either a "real" AI support when AI output is teacher verified and used to assist students to overcome barriers to tasks, or as a "pants on fire" when students are using unverified AI output. Based on these findings, Critical AI Literacy in inclusive EFL contexts might be better interpreted not as a general attitude towards AI, but as a verifiable practice to teach.