📄 social_science

Revisiting AI-Driven Speaking Practice and EAP Replication Study among Turkish Learners

This exploratory replication study demonstrates that Turkish EAP learners who engaged in four weeks of AI chatbot practice showed measurable improvements in oral fluency, reduced speaking anxiety, and increased learner independence, with acoustic and qualitative data suggesting early proceduralization of language skills and a shift toward more ambitious linguistic output.

Original authors: alireza navidmoghaddam gavgani

Published 2026-07-07

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: alireza navidmoghaddam gavgani

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are learning to play a new instrument, like the piano. In a traditional classroom, you practice in front of a teacher and your classmates. If you hit a wrong note, you might feel embarrassed, your heart might race, and you might freeze up. This is very similar to how many Turkish university students feel when they have to speak English in class. They are afraid of making mistakes in front of others, which makes it hard for them to speak smoothly.

This research paper is like a story about a group of ten Turkish students who tried a different kind of practice: talking to a friendly, non-judgmental robot (an AI chatbot called "EAP Talk") instead of a human teacher.

Here is the simple breakdown of what happened and what they found:

The Experiment: A Four-Week "Robot Gym"

The students went to a computer lab for four weeks. Instead of a human teacher, they had nine sessions with this AI chatbot. The robot acted like a conversation partner who never laughed, never got annoyed, and never gave them a bad grade.

Before they started, and after they finished, the students had to record themselves talking for a minute or two about a topic (like describing a favorite object or a social media app). Researchers used special software (like a high-tech microscope for sound) to listen to how they spoke, not just what they said.

The Three Big Changes

The researchers looked at three specific things to see if the students got better at "flowing" while speaking:

1. The "Talking Time" Increased (The Speaking Ratio)

Before: The students talked for about 94 seconds on average.
After: They talked for about 105 seconds.
The Analogy: Imagine a runner who starts by jogging for 10 minutes, then stops to catch their breath for a minute. After training, they jog for 11 minutes before needing a break. The students didn't just talk faster; they managed to keep the "stream" of their voice going for longer without stopping as much.

2. The "Self-Corrections" Decreased (Repairs)

Before: Students often stumbled, repeated words, or fixed their own sentences about 5.8 times every minute.
After: This dropped to 3.6 times per minute.
The Analogy: Think of a writer who keeps erasing and rewriting the same sentence. At first, they were erasing a lot because they were unsure of the rules. By the end, they were writing more smoothly, needing to erase less. They were getting more confident in their "muscle memory" for English.

3. The "Pauses" Moved to the Right Spots
This is the most interesting part.

Before: Students would pause in the middle of a sentence, like a car stopping in the middle of a turn. This usually meant they were struggling to find the next word.
After: Students started pausing only at the end of complete thoughts or sentences, like a car stopping at a red light or a stop sign.
The Analogy: Imagine a train. Before, the train was stopping randomly on the tracks because it was confused. After the training, the train only stopped at the stations (the natural breaks in the sentence). This shows the students were thinking ahead about their whole story, rather than just scrambling for the next word.

Why Did This Happen?

The paper suggests two main reasons:

The "No-Pressure" Zone: Because the robot couldn't judge them, the students felt less nervous. It's like practicing a speech in an empty room versus in front of a crowd. Without the fear of embarrassment, their brains could focus on speaking instead of worrying.
Practice Makes Automatic: The students practiced in short bursts over four weeks (distributed practice). The paper compares this to learning a skill: at first, you have to think hard about every rule (like "where does the verb go?"). After enough practice, your brain stops thinking about the rules and just "does" the action automatically. The robot helped them make that switch from "thinking hard" to "flowing."

What About the "Silence"?

You might think that if students paused more often (which they did), they were getting worse. But the researchers say the opposite. The students started taking longer pauses, but they took them at the right times (between ideas).

The Metaphor: It's like a chef chopping vegetables. A slow, deliberate chop between ingredients is a sign of a chef who is carefully planning the next step. A chop in the middle of a vegetable is a mistake. The students' pauses became "deliberate planning" rather than "mistakes."

The Limits of the Story

The paper is very honest about what it doesn't prove:

Small Group: Only 10 students were involved, so we can't say this works for everyone yet.
No Human Control: They didn't compare the robot group to a group that just practiced with a human teacher, so we don't know if the robot is better than a human, just that it helped.
Short Time: The study only lasted four weeks. We don't know if the students stayed this good a year later.
Robot's Feedback: The students mentioned the robot sometimes gave generic praise ("Good job!") instead of specific corrections. It was good for building confidence, but maybe not perfect for fixing every grammar mistake.

The Bottom Line

This study suggests that for Turkish students who are nervous about speaking English, talking to a friendly AI robot for a few weeks can help them speak more smoothly, make fewer mistakes, and feel less scared. It helped them move from "stumbling through sentences" to "planning their thoughts and pausing at the right times." It's a promising first step, like a training wheel that helps a rider find their balance before they ride on their own.