Recent advances in AI technologies have attracted significant attention from L2 researchers and educators for their pedagogical potential (Law, 2024). As AI tools have become increasingly accessible and sophisticated, they are being incorporated into L2 instructional contexts to support learners’ language development and classroom engagement. In particular, AI has been explored in a range of pedagogical roles, including speaking partner, feedback provider, language tutor, evaluator, writing assistant, and content generator (Lee et al., 2025).

To contribute to the growing body of empirical research, this talk presents two classroom-based studies that illustrate possible applications of AI tools in EFL instructional contexts. The first study examines AI as an asynchronous feedback provider in English writing in an EAP classroom, focusing on how learners process AI-generated feedback and how such engagement relates to the quality of written outcomes. The second study investigates AI as a conversational partner delivering corrective feedback during text-based interactions, with a focus on the role of feedback timing in grammar learning among L2 English learners.

Together, these studies offer insights into both the potential and the challenges of using AI tools in English language classrooms. The talk concludes with pedagogical implications and directions for future research in classroom contexts.

Artificial Intelligence (AI) is often perceived as a mysterious ‘black box’ that powers tools like ChatGPT and Gemini. However, its impact extends far beyond simple text generation, particularly in the realm of modern engineering. This presentation introduces the core concepts of Deep Reinforcement Learning (RL), a field of AI where machines learn to make optimal decisions through a process of trial and error, much like how humans learn from their mistakes.

The talk begins by explaining the mechanics of RL widely used in the field of Wireless Communication Systems within engineering. The explanation is tailored to educators, demonstrating how AI agents interact with their environment to maximize rewards, and how this differs from other AI training methods, such as supervised or unsupervised learning. To enhance the audience’s understanding, I will share specific examples from my current research in Wireless Communication Systems, illustrating how RL algorithms are used to optimize complex networks and empower autonomous systems to adapt to unpredictable environments.

By shifting the focus from abstract mathematics to real-world applications, this talk aims to provide professionals in the field of education with a clear understanding of how AI learns to solve problems, offering insight into how the principle of “success through failure” is shaping the future of communication technology.