Exploring the Potential of Smart Mirror Voice Chatbot on Young L2 Learners’ Spoken
Interaction
This study examines the potential of combining smart mirror technology
with an AI voice chatbot as an engaging English learning tool for young
children. It also evaluates whether this AI-based language program can
encourage meaningful spoken interactions among young EFL (English as a
Foreign Language) learners. The primary research question focuses on whether
the smart mirror chatbot program can provide an effective and sustainable
conversational solution for EFL learners with limited English exposure. Three
sub-questions are also explored: how seven-year-old children interacted with
the smart mirror chatbot over a 10-week study, the extent to which chatbot use
supported children’s speaking skills, and the factors that facilitated or hindered
ii
the use of smart mirror and task-based chatbot technologies for young learners.
This study involved 18 seven-year-old children over ten weeks. For the
first nine weeks, the chatbot was used in teacher-guided speaking activities in
class, and in the final week, children independently chose to use the chatbot
during free activity time. Data were collected through pre- and post-speaking
tests, chatbot conversation logs, observations, and interviews with teachers and
students. The analysis primarily used a qualitative case study method,
supported by quantitative statistical analysis.
Key findings from the children’s interactions with the smart mirror chatbot
include the following: Children recorded an average of 9.85 conversational
turns in free conversations, significantly higher than the typical 3-5 turns
observed in classroom activities. The children used primarily content words,
asked various questions, and increasingly engaged in diverse and realistic
dialogue. They often used question words and worked to form phrases and
sentences. Additionally, as children used the smart mirror chatbot, they
experienced a state of flow, which aligns with Nakamura and
Csikszentmihalyi’s (2002) characteristics of flow: Integration of Action and
Awareness, Autotelic experience, and an Altered Sense of Time.
In terms of English-speaking skills, both qualitative and quantitative
analyses showed improvements in vocabulary and speaking skills. To assess
whether participants used more words in conversations with the chatbot
depending on the duration or method of chatbot use, I analyzed the number of
words in their utterances weekly, comparing in-class activity periods and free
activity periods. The total number of utterances each week was divided by the
number of speeches turns to calculate the average number of words per turn.
Additionally, common errors in pronunciation and cognition were
identified. During free activity, children preferred using the chatbot with peers,
particularly for problem-solving tasks, leading to frequent peer interactions
where they exchanged opinions and collaborated to complete tasks.
Three main factors facilitated smart mirror chatbot use. First, the chatbot
"Kai" was designed to match children’s developmental levels, enhancing their
engagement. Second, a properly structured curriculum supported chatbot use
by incorporating the flow conditions of challenge-skill balance, clear goals,
and feedback, as per Nakamura and Csikszentmihalyi (2002). Notably,
children engaged in decision-making and meaning negotiation, rather than
merely responding to predefined prompts. Third, the smart mirror itself played
an important role: the large 32-inch screen encouraged group participation,
creating an environment where children could share ideas and interact. Unlike
smaller devices, the mirror allowed children to communicate easily with peers
without shifting their gaze, and many enjoyed watching their reflection, which boosted their confidence in speaking English. Some children even practiced
pronunciation while looking at their reflection in the mirror.
Two challenges emerged: initial resistance from teachers and stakeholders
owing to unfamiliarity with the technology, which subsided over time as they
observed its benefits, and the height of the smart mirror, which required shorter
children to stand on tiptoe. Adjustments to the mirror’s height will be
considered in future research.
The educational implications of this study are noteworthy. This research
is unique in combining a researcher-designed smart mirror with an AI chatbot
developed specifically for seven-year-olds, marking a distinctive and original
contribution to the field.