Socratic and dialogic tutoring AI tutors and learning outcomes Voice and learning out loud
Young children who read with an AI that asked questions understood stories better than those reading with a parent
Children aged 5 to 8 learning English as a foreign language read an interactive e-book either with a built-in AI conversational partner that asked questions and gave support, or with a parent. The AI group did better on reading comprehension, and the two groups did about equally well on vocabulary and story retelling, both right away and in a delayed test. Children were more behaviourally engaged with the AI, but showed more emotional engagement and talked more about the story with their parents.
| Paper | Parent-led vs. AI-guided dialogic reading: Evidence from a randomized controlled trial in children's e-book context |
| Authors | Feiwen Xiao, Ellen Wenting Zou, Jiaju Lin, Zhaohui Li, Dandan Yang |
| Published | 18 July 2025 · British Journal of Educational Technology 56(5), 1784–1813 |
| Study type | Randomized controlled trial, two arms · Peer-reviewed |
| What we read | Abstract only. We read the abstract only; the publisher's PDF was not reachable from our tools, so sample details and numbers beyond the abstract are not included. |
| Summarized |
What they did
The researchers built an e-book with a large language model reading partner that engaged children in dialogic reading, the style of shared reading where the adult asks questions and builds on the child's answers. Children were randomly assigned to read with the AI or with a parent, and were tested on comprehension, vocabulary and retelling immediately and after a delay. Engagement was also observed.
Who took part
67 children aged 5 to 8 learning English as a foreign language.
What they found
- Children in the AI-guided group outperformed the parent-led group in reading comprehension, immediately and in delayed tests (effect sizes not given in the abstract).
- Vocabulary learning and story retelling were comparable between the groups.
- Children reading with the AI showed more behavioural engagement and visual attention.
- Children reading with a parent showed more affective (emotional) engagement and more talk relevant to the story.
Effect size: Not given in the abstract.
Limitations
- We read only the abstract, so we cannot check the study's methods or effect sizes.
- Small sample of 67 children, all learning English as a foreign language.
- The authors built the AI reading partner they evaluated.
- The children were 5 to 8 and the task was reading an e-book on a screen.
What this means for learners and parents
- An AI that asks questions about a story, rather than just reading it out, can support comprehension for young children.
- Parents still brought something the AI did not: warmth and real conversation about the story. The two can work together.
- For parents reading at home, the useful part is the questioning style: ask what might happen next, why a character did something, and build on the answer.
Put it into practice (free tools)
Try it: Ask Poppy is voice-first homework help for kids, built first for ages 7 to 10 (teens and adults use it too), on iPhone, iPad and Mac. It asks guiding questions instead of handing over the answer, so children learn how to find it. Free with 30 messages a day, then $3.99/week or $9.99/month. (We make it.)
Citation
Feiwen Xiao, Ellen Wenting Zou, Jiaju Lin, Zhaohui Li, Dandan Yang (2025). Parent-led vs. AI-guided dialogic reading: Evidence from a randomized controlled trial in children's e-book context. British Journal of Educational Technology 56(5), 1784–1813. https://doi.org/10.1111/bjet.13615
Related summaries
- Unrestricted ChatGPT help raised practice scores but lowered exam scores · Hamsa Bastani et al., 2025 · Preregistered randomized controlled trial
- A carefully designed AI tutor beat an active-learning physics class on a short-term test · Greg Kestin et al., 2025 · Randomized controlled trial with a crossover design
- Supervised AI tutoring matched human tutors in UK maths classrooms · LearnLM Team (Google) et al., 2025 · Exploratory randomized controlled trial
- An AI coach for human tutors raised topic mastery, most for weaker tutors · Rose Wang et al., 2024 · Preregistered randomized controlled trial