Lecturers’ Experiences in Human Expertise and Generative AI in Developing Communicative Speaking Materials for English Foreign Language (EFL)

In classrooms around the world, teachers face a constant challenge: how to help students learn to speak a new language not just correctly, but meaningfully. For decades, the most trusted approach has been to focus on real communication rather than memorizing rules. This method, known as Communicative Language Teaching, asks students to use language to solve problems, share ideas, and interact with one another, much like they would in daily life. The goal is to build a skill where a student can navigate a conversation, not just recite a grammar chart. Today, a new force has entered this classroom dynamic: generative artificial intelligence. These are computer systems capable of writing text, creating scenarios, and drafting lesson plans in seconds. While this technology promises to save teachers time and spark new ideas, it also raises a quiet but urgent question. If a machine can write a lesson plan instantly, does the teacher still need to be the expert designer of that lesson? The answer matters because language learning is deeply human; it relies on understanding culture, emotion, and the messy, unpredictable nature of real conversation.

A team of researchers set out to watch how English teachers in Indonesia are navigating this new reality. They focused on five university lecturers who teach speaking skills to students training to become English teachers. These educators use tools like ChatGPT and similar programs to help them create materials for their classes. The researchers did not ask if the teachers used the technology; they asked how they used it. Through detailed conversations, the team listened to how these lecturers balanced their own professional experience with the instant output of artificial intelligence. What they found was a clear pattern of human control. The teachers did not simply copy and paste what the computer gave them. Instead, they treated the artificial intelligence as a helpful assistant that could brainstorm ideas or draft a first version of a lesson, but they insisted on being the final editors and decision-makers.

The lecturers explained that their expertise lies in knowing their specific students and the local context. They described a process where they might ask an artificial intelligence to suggest a role-play activity about a market or a debate topic. The computer would quickly generate a script or a set of instructions. However, the teachers then stepped in to reshape that content. They noticed that the computer often produced conversations that were too perfect, too polite, and too logical. Real human speech, they pointed out, is full of interruptions, hesitations, and moments where people misunderstand each other before finding a way to clarify. The lecturers would rewrite the computer's smooth dialogue to include these realistic imperfections, ensuring the students practiced for the actual world, not a polished simulation. They also adjusted the topics to fit their local environment, swapping generic examples for issues relevant to their region, such as local tourism or environmental concerns in South Sulawesi.

This careful editing was not just about fixing errors; it was about protecting the educational value of the lesson. The teachers found that while artificial intelligence could save hours of preparation time, it could not replace the judgment needed to decide what a student actually needed to learn. One lecturer noted that the computer might suggest a difficult activity that looked good on paper but would confuse the specific group of students in front of them. Another mentioned that the computer often missed cultural nuances, offering examples that felt foreign or inappropriate to Indonesian learners. The teachers acted as a filter, ensuring that every piece of material was culturally relevant and pedagogically sound. They viewed the technology as a tool to overcome creative fatigue, offering a spark of new ideas, but they remained the ones who decided which ideas to keep and how to refine them.

The study also highlighted how these teachers were preparing their own students for this new digital age. Rather than banning artificial intelligence, the lecturers encouraged their students to use it critically. They asked students to compare what the computer wrote with what they knew from textbooks or real-life experience. The goal was to teach students to question the output, to spot biases, and to verify facts. The teachers emphasized that the machine is not neutral; it carries its own assumptions and patterns. By teaching students to evaluate the technology, the lecturers were ensuring that the next generation of teachers would not just use tools, but understand them. This approach turned the classroom into a space for critical thinking, where the technology served as a subject of study as much as a helper.

Ultimately, the research suggests that the most effective way to use generative artificial intelligence in language teaching is to keep the human teacher firmly in charge. The technology excels at speed and variety, offering a wide range of drafts and ideas in moments. But the human teacher brings the essential elements of context, empathy, and professional judgment. The lecturers in this study did not see themselves as being replaced by machines; they saw themselves as the conductors of a new orchestra, using the technology to expand their resources while ensuring the music remained true to the needs of their students. The balance they struck was not about how much work the computer did, but about who made the important decisions. As long as the teacher remained the one defining the learning goals, adapting the materials to the students, and judging the quality of the interaction, the technology served as a powerful support. The study concludes that the future of language teaching does not lie in choosing between human expertise and artificial intelligence, but in finding a partnership where the machine handles the routine, and the human mind handles the meaning.