People · Practice · Possibility
Independent research for education leaders
Follow a question across the research. Search findings, sources, and the implications for your schools.
24 stories · Clear filters
Which parts of lesson preparation should teachers keep? Researchers followed 90 mathematics teachers in 18 Chinese secondary schools, coding 897 lessons and comparing plans with AI interaction logs. When a teacher delegated more of the questioning and…
Can general-purpose AI improve what students remember after assistance ends? In an experiment with 211 undergraduates, students studied an unfamiliar technical topic and wrote an essay within a fixed period, with or without permission to use AI. All knowledge…
A high school capstone can involve dozens of different internships, making individualized weekly feedback difficult to sustain. This design report describes a chatbot configured around course rubrics, requirements and learning theories to help seniors set…
A model that solves mathematics well may still explain it poorly. Sherpa trains a tutoring model through repeated conversations with simulated learners that require different instructional approaches. Across those artificial learner types, reported…
What support do teachers need to learn AI well? A survey of 213 Turkish teachers examined planning, monitoring, help-seeking, perceived institutional support and self-evaluation. An exploratory analysis grouped 173 respondents with broadly stronger learning…
Does an integrated AI teaching system improve computational thinking? A 14-week pilot compared two university classes totaling 120 students. One class used a package combining learning resources, adaptive sequencing, interactive support and feedback; the…
Which classroom relationships should AI support? A mini-review brings together 32 studies on educator communication and AI-mediated higher education. It finds that teacher clarity, rapport and responsiveness remain associated with student outcomes, while…
Support for AI and concern about its effects can coexist in the same educator. A survey of 270 teachers across seven campuses of a Pakistani private university asked about pedagogical benefits and structural risks. Teachers rated interactivity and assessment…
Can children explain when a machine needs to learn from examples? In one Turkish primary school, 93 third-graders took part in an eight-week comparison. Two classes built LEGO-based systems that contrasted trained pose classifiers with fixed rules; two…
An answer can be correct without revealing how a child thinks. Researchers tested a chatbot that asked 33 fifth- and sixth-graders to explain and compare ratio-solving strategies after a short lesson and assessment. Built with GPT-4o through Playlab, Math…
Can a questioning partner help future teachers build stronger scientific arguments? An eight-week study compared 124 biology teacher candidates in two intact classes at an Uzbek university. One group used a GPT-4-based Socratic protocol that repeatedly…
A wrong final answer can hide several successful intermediate steps. This study turns mathematical tasks inside tutoring conversations into standalone questions, then combines those responses with ordinary question records to estimate difficulty and student…
A tutor that withholds the answer can still lead a learner in the wrong direction. SocraticRepair fine-tunes a small programming model to guide Python debugging through explicit teaching plans. In 159 simulated interactions per system, the verified-plan…
Private help may make it easier for students to ask questions, but that is different from learning more. Across four Chinese universities, an EduBot study compared eight course sections during a semester, with 555 students completing three survey waves. The…
Bilingual access requires more than translating a fluent answer. This prototype translates Mandarin questions into English, retrieves relevant course materials, and produces bilingual guidance with a technical glossary. Researchers tested 91 selected posts…
What would it take for students without programming backgrounds to examine a published AI model rather than merely use its predictions? PRAISE describes a workflow for selecting, installing, reproducing and validating scientific models, with explicit routes…
When should an AI tutor hand a struggling learner back to a teacher? A six-week algebra pilot with 32 eighth-grade boys combined a rule-based practice system, a generative AI questioning partner and direct teacher support. Classroom use exposed practical…
Can building a simple model help children understand AI? Across 11 Madrid-area schools, 996 pupils aged 10-11 completed four sessions using visual programming and tools such as Teachable Machine. They created and trained models rather than only consuming…
What helps kindergarten teachers use AI responsibly? Researchers in Malang, Indonesia, interviewed ten teacher leaders and five administrators, then surveyed 250 kindergarten teachers with prior AI-use experience. Teachers described planning stories,…
Does a structured questioning routine improve how learners work with AI? In a television-production course, 83 undergraduates in three intact classes used DeepSeek for a short-film project. One class asked free-form questions; two followed progressive or…
Receiving comments is different from learning to evaluate and use them. A ten-week study assigned three intact university English classes to AI, peer or teacher feedback, with 126 students completing the analysis. AI and peer groups improved their reported…
A paper-first routine asks students to map their argument before consulting AI. Over two years and nine design cycles, a South African university project developed this approach with a six-person design team and a cohort of 150 foundation-level students.…
Human oversight is often promised more clearly than it is evaluated. A review of 78 higher-education studies examined educators' use of generative AI to design instruction, the competencies involved and professional learning approaches. Design-stage activity…
Students can value AI while still wanting clearer rules. Research in Ecuador combined interviews with 14 lecturers and eight students, a 200-student survey and institutional document review. Participants described useful assistance with academic work…