🏷️ intelligent-tutoring

> **Synthesis:** Udeshi et al. (2026), the team behind **Khanmigo** (Khan Academy's K-12 AI tutor, launched 2023), describe the metrics they use to measure AI tutoring quality and student engagement, …

2026-08-13 · ai-tutoring, k-12, llm, personalized-learning, efficacy-study

> **Synthesis:** Niousha, Kang, & Norouzi (2026) introduce **INTERNAL STUDENT DIALOGUE (INSIDE)**, a student modeling framework that fine-tunes LLMs to both *act* like students and *think* like them. …

2026-08-13 · student-modeling, llm, learning-analytics, personalized-learning, ai-education

> Ye et al. (2026) introduce **AgentSchool**, an LLM-driven multi-agent [[simulating-students|simulator]] that models learning as **state transition rather than prompted behavior**. It couples cogniti…

2026-08-12 · simulating-students, agentic-ai, multi-agent, adaptive-learning, zone-of-proximal-development

> Pardos & Bhandari (2024) report a randomized efficacy study (N=274) comparing ChatGPT-generated hints to human tutor-authored hints and a no-help control across four mathematics subject areas. Only …

2026-08-12 · generative-ai, llm, ai-tutoring, scaffolding, math-education

> OATutor (Open Adaptive Tutor) is the first open-source adaptive tutoring system built on Intelligent Tutoring System (ITS) principles, developed at UC Berkeley's CAHL Lab. It combines an MIT-license…

2026-08-12 · adaptive-learning, open-source, knowledge-tracing, math-education, higher-ed

> **Cognitive diagnosis** — the inference of a learner's latent knowledge state — the specific concepts, skills, and misconceptions they have or lack — from their responses or behavior. It is the asse…

2026-08-12 · student-modeling, educational-measurement, knowledge-tracing, assessment, learning-analytics

> **Simulating students** — using LLM-based agents to model learner behavior, cognition, and social dynamics for educational research, design, and training. Simulated students let researchers evaluate…

2026-08-12 · simulation, student-modeling, knowledge-tracing, cognitive-diagnosis, agentic-ai

> **Adaptive learning** — AI-driven educational systems that adjust content, pacing, and instructional strategies based on individual learner characteristics and performance. Adaptive learning is the …

2026-08-09 · knowledge-tracing, personalized-learning, student-modeling, scaffolding, cognitive-diagnosis

> **Math Education** — the study of how students learn mathematics and how AI can support mathematics teaching, spanning affective tutoring, cognitive diagnosis from handwritten work, productive strug…

2026-08-09 · math-education, stem-education, ai-tutoring, k-12, generative-ai

> **Scaffolding** — structured support that helps learners accomplish tasks they cannot yet complete independently, with support fading as competence grows. In AI in education, scaffolding is the prim…

2026-08-09 · zone-of-proximal-development, over-reliance, cognitive-load-theory, feedback-loop, ai-literacy

> **Socratic Method** — a pedagogical approach rooted in guided questioning and dialogue rather than direct instruction, now being adapted for generative AI tutoring systems. In AI in education, the S…

2026-08-09 · ai-education, generative-ai, higher-ed, student-experience, scaffolding

> **Student modeling** — the broad practice of representing learner characteristics including knowledge, skills, affective states, engagement, and preferences in computational form. Student modeling i…

2026-08-09 · student-modeling, knowledge-tracing, adaptive-learning, personalized-learning, learning-analytics

> **Synthesis:** Graf et al. (2026) transformed a conventional Learning Analytics Dashboard (LAD) into an interactive ILAD by adding an LLM-powered pedagogical agent and a Judgement of Learning (JoL) …

2026-08-08 · learning-analytics, metacognition, llm, higher-ed, engagement-metrics

> **Synthesis:** Zhang et al. (2026) introduce TutorMoments, a replay-based evaluation framework that tests whether LM tutors adapt their pedagogical actions to context — scaffolding when support is n…

2026-08-08 · scaffolding, llm, llm-evaluation, k-12, math-education

> **Synthesis**: Pedagogical agents are AI-driven conversational interfaces embedded in learning environments that use pedagogical strategies (eliciting, telling, scaffolding) to support learner engag…

2026-08-08 · llm, student-ai-interaction, scaffolding, higher-ed, generative-ai

> **Synthesis:** A three-semester, 999-student analysis of hint usage in a K-12 mathematics ITS finds that two simple, interpretable indicators—premature hint requests and superficial hint reading—are…

2026-08-06 · learning-analytics, hint-systems, k-12, math-education, gaming-the-system

> **Help-Seeking** — a key concept in AI in education research. Explored across 4 articles in this wiki.…

2026-08-06 · help-seeking, student-experience, metacognition, higher-ed, llm

> **EduClaw-Bench: A Long-Horizon Benchmark for Pedagogical LLM Agents with Simulated Learners** — Introduces a 30-day long-horizon benchmark for pedagogical LLM agents using simulated learners ground…

2026-08-05 · llm, agentic-ai, benchmark, knowledge-tracing, student-modeling

> **A fully open-source agentic tutoring framework that closes the loop between citation-grounded problem tutoring and difficulty-calibrated question generation**, powered by a hybrid personalization …

2026-08-04 · llm, agentic-ai, personalized-learning, rag, question-generation

> **A new paradigm for online education replacing MOOCs with LLM-driven multi-agent AI classrooms**, piloted at Tsinghua University with 100K+ learning records from 500+ students. MAIC uses specialize…

2026-08-04 · llm, agentic-ai, multi-agent, mooc, online-education

> Robinson, Gormley, Ribeiro & Loeb (2026) ran two RCTs showing that AI tutoring's binding constraint is **take-up, not capability**: despite dedicated session time, nearly half of students never used…

2026-08-03 · equity, engagement-metrics, k-12, ai-literacy, student-experience

Presents **StructRAG**, a pattern-aware framework that improves how AI tutoring systems interpret **complex engineering diagrams** (circuit schematics, network topologies, block flowcharts) in STEM. C…

2026-08-03 · stem-education, llm, generative-ai, automated-grading, feedback-loop

> **Authors:** Viola Deutscher, Herbert Thomann, Olga Zlatkin-Troitschanskaia, Ulrike Weyland, Stephan Abele, Amory H. Danek, Samuel Greiff, Andreas Rausch, Susan Seeber, Jürgen Seifried, Esther Winth…

2026-08-01 · professional-training, efficacy-study, generative-ai, adaptive-learning, simulation-based-learning

> **Agentic AI** — AI systems that autonomously plan, execute, and adapt multi-step workflows to achieve learning goals, going beyond single-turn Q&A to act as persistent, goal-directed collaborators:…

2026-08-01 · agentic-ai, generative-ai, llm, higher-ed, scaffolding

> **AI tutoring** — the use of AI (especially [[llm|LLMs]] and [[intelligent-tutoring|intelligent tutoring systems]]) to provide personalized, adaptive, scalable instructional support: conversational …

2026-08-01 · generative-ai, llm, scaffolding, feedback-loop, higher-ed

Pre-registered study auditing whether general-purpose helpfulness rubrics can distinguish direct answer-giving from pedagogical guidance in LLM tutors. Uses deterministic detectors for answer leakage …

2026-07-31 · llm, automated-grading, benchmark, feedback-loop, adaptive-learning

Published in *Computers and Education: Artificial Intelligence*, accepted 27 July 2026. 📄 doi:10.1016/j.caeai.2026.100653…

2026-07-29 · agentic-ai, ai-literacy, educational-theory, generative-ai, higher-ed

> **Critical AI Tutors: Empower or Enslave?** — A position paper presented at the AIED 2025 workshop that issues a stark warning: unchecked use of AI tutors risks creating a generation of cognitively …

2026-07-29 · critical-thinking, over-reliance, ai-assistance-reduces-persistence, student-experience, regulation

> **Synthesis:** A custom Socratic AI chatbot deployed in a large-enrollment introductory mechanics course with 150 first-year STEM majors, demonstrating that AI-driven Socratic dialogue can foster ex…

2026-07-29 · socratic-method, physics-education, generative-ai, socratic-questioning, computational-thinking

> **MathBuddy: Affective Math Tutoring** — EMNLP 2025 Demo. An emotionally aware LLM-powered mathematics tutor that dynamically models student emotions from both conversational text and facial express…

2026-07-29 · math-education, affective-computing, k-12, pedagogical-llm-training, affective-tutoring

> **PersonaVLM** introduces an agent framework for long-term personalization of multimodal LLMs, enabling AI tutors to remember, reason about, and align with a learner's evolving preferences across hu…

2026-07-29 · personalized-learning, student-modeling, llm, ai-tutoring, affective-tutoring

> **The Multimodal Interference Effect** describes a systemic accuracy drop when LLMs encounter image-rich STEM problems: from ~96% on text-only physics problems to ~74% on multimodal ones. A simple t…

2026-07-29 · multimodal, socratic-method, stem-education, llm, ai-tutoring

> **Comprehensive Review of Intelligent Tutoring Systems** — Journal of Computers in Education (2025). A systematic literature review covering 2010–2025 that analyzes the deployment and effectiveness …

2026-07-29 · literature-review, adaptive-learning, pedagogical-llm-training, educational-nlp, socratic-questioning

Brisson, Segarra and Smits present a didactically-driven LLM teacher assistant for a university dimensional modeling (data warehousing) course. Unlike most educational chatbots that delegate pedagogic…

2026-07-28 · llm, higher-ed, pedagogy, rag, scaffolding

> **Affective computing** in education uses physiological and behavioral signals to sense learner emotion and adapt instruction — see [[affective-text-wearable-student-health]], [[multimodal-affective…

2026-07-28 · student-experience, higher-ed, llm, generative-ai, learning-analytics

> **Open-source** AI in education is studied in [[lata-ferpa-compliant-local-llm-autograder]], [[vismatic-secure-sandbox-cs-education]], and [[open-source]] (tag) pages: local open models address [[pr…

2026-07-28 · llm, stem-education, educational-theory, generative-ai, ai-education

> **Reinforcement learning** trains AI tutors and agents through reward signals: [[special-r1-rl-special-education]], [[singh-eduqwen-pedagogical-rl-2026]], [[pedagogical-safety-rl]], and [[ai-coachin…

2026-07-28 · llm, pedagogical-safety, special-education, personalized-learning, k-12

Kutti AI addresses a persistent equity gap in educational technology: nearly all edtech assumes a visual interface, excluding an estimated 1.4 million blind children worldwide. The system inverts this…

2026-07-27 · adaptive-learning, special-education, k-12, equity, personalized-learning

EduGuard is a retrieval-augmented generation (RAG) tutoring framework that directly confronts the safety and pedagogical failures of unrestricted LLM tutors in introductory programming. Unrestricted t…

2026-07-20 · llm, generative-ai, stem-education, over-reliance, hallucination-risk

This landmark field experiment is among the first randomized controlled trials to causally demonstrate that **unguarded generative-AI tutoring can harm skill acquisition**, not merely fail to help. Co…

2026-07-19 · generative-ai, over-reliance, stem-education, k-12, rct

LEA (Learning Engagement Assistant) is an **agentic AI tutoring system** that couples course-specific retrieval-augmented generation (RAG) with structured [[knowledge-tracing]] / Knowledge Component (…

2026-07-16 · llm, generative-ai, higher-ed, stem-education, formative-assessment

Addresses the lag between LLM integration into K-12/higher education and reliable methods for evaluating pedagogical quality. The authors introduce a knowledge-distillation approach to automate AI-tut…

2026-07-14 · automated-grading, llm, higher-ed, k-12

STEM education faces challenges in personalization and interdisciplinary integration. AI technology has brought new possibilities, but the mechanisms by which AI reshapes the STEM education ecosystem …

2026-07-09 · stem-education, generative-ai, scaffolding, adaptive-learning, learning-analytics

Introductory programming instruction relies on hands-on practice and short learning activities to support mastery of foundational concepts. Although many such learning resources exist, organizing and …

2026-07-09 · knowledge-tracing, personalized-learning, adaptive-learning, programming-its, rag

Deploying LLM tutors in K-12 raises concerns around privacy, cost, and reliance on proprietary models, motivating small language models (SLMs) as an alternative. The authors introduce **CSTutorBench**…

2026-07-08 · llm, k-12, privacy, benchmark, cs-education

Prompt engineering is a critical yet undertaught skill for software developers, poorly served by traditional instruction because of its evolving, interactive, context-dependent nature. The authors int…

2026-07-08 · agentic-ai, cs-education, feedback-loop, formative-assessment, llm

The rapid integration of [[llm]]s into [[intelligent-tutoring]] threatens to reduce mathematical learning to mere answer generation. This paper presents a design framework for AI tutors that act as re…

2026-07-03 · llm, scaffolding, formative-assessment, higher-ed, automated-grading

> **Lorenzo Stacchio, Michele Giordano, Daniele Berardini, Primo Zingaretti, Emanuele Frontoni** — submitted 17 Jun 2026…

2026-07-02 · generative-ai, k-12, equity, personalized-learning, llm-in-education

This survey provides the first systematic review of automated presentation coaching systems, organizing them along a five-dimensional task taxonomy: segmental pronunciation, lexical stress, suprasegme…

2026-06-29 · llm, feedback-loop, professional-training, language-learning, personalized-learning

This paper explores how an embodied AI agent can act as a [[scaffolding|coach]] that accelerates human motor-skill development using [[adaptive-learning|reinforcement learning]]. The authors argue tha…

2026-06-25 · personalized-learning, scaffolding, adaptive-learning, professional-training, generative-ai

SupplyNet is a gamified visual simulation system that uses a contextual graph-based [[llm]] multi-agent framework to model interdependent supply chain dynamics. Designed for [[professional-training]] …

2026-06-24 · llm, generative-ai, active-learning, professional-training, simulation-based-learning

Scheinberg et al. (2026) analyze 78 university learners of German across four sites completing a counterbalanced spot-the-difference task with both a human peer and a real-time AI partner. Using diari…

2026-06-23 · language-learning, llm, student-experience, personalized-learning

Castillo Ventura et al. (2026) address the gap between instructional design of digital mathematics resources and their pedagogical use in classrooms. Their work-in-progress framework translates learni…

2026-06-23 · teacher-role, k-12, stem-education, scaffolding, ai-literacy

Ganganath et al. (2026) introduce CURIOBOT, a framework that operationalizes Berlyne's four collative variables (novelty, complexity, conflict, uncertainty) as adaptive linguistic interventions in con…

2026-06-23 · llm, metacognition, scaffolding, active-learning, self-regulated-learning

Park et al. (2026) explore AI-powered automated feedback for tutors on Ringle, a popular online English tutoring platform in the gig economy. Their research probe analyzed tutors' lessons and provided…

2026-06-23 · feedback-loop, higher-ed, language-learning, teacher-role

> **Knowledge tracing** — modeling what learners know over time by tracking their performance on exercises and predicting future mastery. It is the wiki's richest modeling thread, spanning Bayesian, d…

2026-06-23 · knowledge-tracing, learning-analytics, adaptive-learning, student-modeling, personalized-learning

> **Po-Chin Chang, Nicholas Hogan, Aske Plaat, Michiel T. van der Meer** (2026). arXiv cs.AI preprint…

2026-06-19 · llm, k-12, personalized-learning, scaffolding, adaptive-learning

AI-driven assessment of human tutor training performance correlates with real-life tutoring quality; bridges the gap between training metrics and classroom practice. AI-Driven Assessment of Human Tuto…

2026-06-18 · automated-grading, feedback-loop, teacher-role, efficacy-study, adaptive-virtual-patient-psychotherapy-training

ASTRA uses autonomous AI sim-pilots for scalable air traffic control training; reduces dependency on human role-players while maintaining realistic scenario complexity. ASTRA: A Scalable Next-Generati…

2026-06-18 · adaptive-learning, professional-training, active-learning, llm, adaptive-virtual-patient-psychotherapy-training

MedEasy multi-agent system simulates standardized patients with varying conditions for medical consultation training; outperforms script-based approaches in realism and adaptability. MedEasy: Designin…

2026-06-18 · generative-ai, higher-ed, active-learning, feedback-loop, adaptive-virtual-patient-psychotherapy-training

> **Jaward Sesay, Yue Yu, Siwei Dong, Yemin Shi, Guangyao Chen, Borje F. Karlsson** (2026). arXiv cs.CL…

2026-06-17 · llm, generative-ai, personalized-learning, higher-ed, pedagogical-llm-training

> **Lan Luo, Anqi Wang, Muzhi Zhou, Junhua Zhu, Jie Cai, Ao Yu, Hui Pan** (2026). arXiv cs.HC…

2026-06-17 · scaffolding, k-12, personalized-learning, llm

> **Alexandra Neagu, Jeffrey T. H. Wong, Marcus Messer, Rhodri Nelson, Peter B. Johnson** (2026). Pluralistic Alignment Workshop @ ICML 2026…

2026-06-17 · scaffolding, llm, benchmark, efficacy-study, student-experience

> **Xinfei Cen, Vincent Aleven, Kenneth R. Koedinger, Conrad Borchers, Paulo F. Carvalho** (2026). EC-TEL 2026…

2026-06-17 · personalized-learning, efficacy-study, higher-ed, student-experience, self-regulated-learning

> **Farnaz Baksh, Matevz B. Zorec, Feiazie Baksh, Karl Kruusamae** (2026). ICSR + ART 2026, London…

2026-06-17 · k-12, higher-ed, student-experience, stem-education, human-ai-collaboration

Team-based projects are a cornerstone of engineering and computing courses, but unstructured team formation often leads to poor project outcomes due to misaligned student interests and inadequate skil…

2026-06-16 · edtech-platform, higher-ed, stem-education, personalized-learning, learning-analytics

> Asks what is gained and lost when 'collaboration' is applied freely to human-AI interaction. Argues true collaboration requires symmetric/negotiated relationship, shared goals, low and shifting divi…

2026-06-16 · scaffolding, active-learning, student-experience, teacher-role, learning-analytics

Studies whether public LLM tutoring benchmarks distinguish learning-supportive behavior from mere answer production. Proposes a lightweight diagnostic based on the gap between solving-oriented and ped…

2026-06-16 · llm, feedback-loop, scaffolding, benchmark, learning-analytics

This paper investigates whether [[llm|large language models]] can serve as scalable proxies for students by simulating realistic logical errors in code submissions. Using the CodeWorkout dataset of 74…

2026-06-15 · llm, stem-education, student-experience, learning-analytics, efficacy-study

**Angela Chen, Siwei Jin, Catherine Bao, Canwen Wang, Robert E. Kraut, Tongshuang Wu, Haiyi Zhu** — cs.CY, cs.HC The Adaptive Virtual Patient (AVP) is an LLM-driven simulated patient for psychotherapy…

2026-06-10 · professional-training, llm, generative-ai, student-experience, higher-ed

**Misan Paul Etchie, Taiwo Olutosin** — cs.CY, cs.AI, cs.HC This paper proposes an AI-integrated LMS designed specifically for middle school instruction, addressing the gap between current LMS platfor…

2026-06-10 · k-12, adaptive-learning, personalized-learning, formative-assessment, edtech-platform

**Oleksandra Poquet, Mani Shankar Nanduri, Maria Ximena Salinas Loyer, Matthias Stadler, Michael Sailer, Jelena Jovanovic** — Accepted at EC-TEL 2026 — cs.HC, cs.ET This study compares two approaches …

2026-06-10 · writing-education, llm, student-experience, higher-ed, ai-literacy

**Jingzhe Lin, Hengbin Yu, Yongdan Zeng, Fangwei Zhong** — ICML 2026 — cs.MA, cs.CY EduMirror introduces a multi-agent simulator for studying educational social dynamics, addressing the dilemma that o…

2026-06-10 · student-experience, llm, higher-ed, scaffolding, generative-ai

**Yifan Liu, Jaime Arguello, Orland Hoeber, Chang Liu et al.** — cs.IR, cs.AI, cs.HC This report summarizes the CHIIR 2026 Workshop on Generative AI and Academic Search (GAI&AS), which examined how Ge…

2026-06-10 · generative-ai, llm, ai-literacy, policy-maker, higher-ed

This paper introduces a pipeline that maps student questions directed at a conversational AI teaching assistant to curriculum topics using a few-shot text classifier, grounded in a GPT-4-extracted pre…

2026-06-10 · knowledge-tracing, llm, student-modeling, higher-ed, learning-analytics

**Yi-Chieh Lee, Nattapat Boonprakong, Yugin Tan, Harold Soh et al.** — Workshop report from NUS-Google Workshops — cs.CY This white paper synthesizes findings from two international NUS-Google Worksho…

2026-06-10 · cs-education, generative-ai, curriculum-design, higher-ed, llm

**Yibo Meng, Ruiqi Chen, Zhiming Liu, Xiaolan Ding** — Accepted at MobileHCI 2026 — cs.HC TibetCPR is a low-cost, self-guided CPR training system that pairs depth-driven electrotactile feedback with r…

2026-06-10 · professional-training, llm, formative-assessment, student-experience, edtech-platform

Reports on cross-boundary Community-Based Learning where undergraduate students develop AI-enabled solutions for cultural heritage preservation and sustainable development. The paper argues that AIED …

2026-06-09 · higher-ed, ai-literacy, equity, student-experience, teacher-role

> **Synthesis:** Recent work in Technology-Enhanced Learning and HumanComputer Interaction highlights the importance of transparency and trust calibration in AI-supported learning environments as they…

2026-06-03 · student-experience, hallucination-risk, trust-calibration, llm, help-seeking

> **Synthesis:** Balancing AI responsibility with privacy, safety, and utility: Unlearning in large language models for mathematics education…

2026-06-03 · llm, privacy, k-12, pedagogical-safety

> Experimental study comparing Guided vs. Unrestricted LLM access. Explicit training in reasoning-focused scaffolding (stepwise hints, verification) led to significantly better independent performance…

2026-06-02 · scaffolding, higher-ed, metacognition, prompt-engineering, agentic-workflows

> **Authors:** Unggi Lee, Jihoi Na, Yeil Jeong, Haeun Park, Yeonju Jang (2026)…

2026-06-01 · llm, special-education, personalized-learning, reinforcement-learning, k-12

> **Authors:** Shim Jaechang, Unggi Lee (2026) — CIKM 2026…

2026-06-01 · llm, benchmark, efficacy-study, automated-grading, formative-assessment

A key part of developing large language model (LLM)-powered, automated tutoring tools is student simulation, i.e., using LLMs to role-play as students, which can facilitate tutor model evaluation and …

2026-05-29 · llm, student-experience, learning-analytics, personalized-learning, reinforcement-learning

The widespread adoption of AI chatbots in education will drastically change learning, making responsible deployment a critical concern. While large language models (LLMs) might have access to sources …

2026-05-29 · llm, scaffolding, teacher-role, over-reliance, learning-analytics

**Catching the Correct Answer Trap** — accepted at AIED 2026 — exposes a critical blind spot in [[intelligent-tutoring]] systems: they systematically fail to detect misconceptions when students arrive…

2026-05-28 · automated-grading, formative-assessment, llm, scaffolding, generative-ai

**KT4EQG: Personalized Exercise Question Generation via Knowledge Tracing** bridges two key AI-in-education paradigms: [[personalized-learning]] through question generation and [[learning-analytics]] …

2026-05-28 · adaptive-learning, automated-grading, learning-analytics, llm, personalized-learning

This is the first empirical study of what happens when AI agents are embedded **persistently** in a real academic research environment — with durable memory, local files, external tools, scheduled rou…

2026-05-27 · llm, generative-ai, higher-ed, agentic-ai, faculty-development

SlidesQAQA is a Flask-based system that extracts text and rendered images from PDF lecture slides and processes them through a four-stage [[llm]] pipeline: **window planning** (segment extraction), **…

2026-05-27 · prompt-engineering, generative-ai, llm, higher-ed, scaffolding

This paper presents a fine-grained taxonomy categorizing novice programmers' difficulties with reading and understanding control flow constructs — specifically conditionals (selection) and loops (iter…

2026-05-27 · cs-education, stem-education, student-experience, higher-ed, benchmark

📄 DOI: 10.32473/flairs.39.1.141554 Codify (also called AI Tutor) is an [[intelligent-tutoring]] system that leverages [[llm|LLMs]], competency tracking, and adaptive assessment to provide Socratic, d…

2026-05-26 · llm, stem-education, higher-ed, adaptive-learning, self-regulated-learning

StanBKT introduces an open-source Python package for Bayesian Knowledge Tracing (BKT) that moves beyond traditional expectation-maximization (EM) point estimates to full Bayesian inference via Stan. T…

2026-05-25 · learning-analytics, adaptive-learning, benchmark, open-source, adaptive-learning-systems

Serious games are widely used for learning and training across domains such as healthcare, defense, and education. This chapter examines how contemporary AI approaches may support real-time instructio…

2026-05-22 · generative-ai, adaptive-learning, edtech-platform, professional-training, llm

**Annie Yuan (2026)**. arXiv preprint (cs.HC). Current AI-driven educational systems primarily rely on behavioural analytics and performance metrics, lacking the ability to model expert cognition used…

2026-05-22 · learning-analytics, adaptive-learning, student-experience, teacher-role, ai-literacy

Intelligent Tutoring Systems often grant learners shared control over skill and problem selection. We propose a simulation-based framework to examine how learner task-selection strategies and system c…

2026-05-22 · mastery-learning, adaptive-learning, engagement-metrics, simulation, k-12

Theodora and Tselios (2026) provide a policy-oriented synthesis of AI's dual role in adult and [[lifelong-learning]] contexts — as both an enabler of personalized, scalable education and a source of s…

2026-05-21 · lifelong-learning, policy-maker, equity, ai-literacy, learning-analytics

Benedetti (2026) introduces a theoretically grounded framework for AI-enabled pedagogical accompaniment that explicitly centers human agency — an approach described as "human-in-command" rather than m…

2026-05-21 · scaffolding, human-in-the-loop, pedagogical-safety, ai-literacy, student-experience

Yang et al. (2026) tackle a fundamental tension in AI-augmented classrooms: how to balance teacher orchestration with student agency during dynamic transitions between individual and collaborative wor…

2026-05-21 · teacher-role, student-experience, k-12, human-in-the-loop, collaborative-learning

📄 DOI: 10.1016/j.caeai.2022.100074 This paper introduces **XAI-ED**, a framework for explainable AI that is purpose-built for education. It argues that while XAI in education shares common ground wit…

2026-05-21 · learning-analytics, bias-mitigation, equity, privacy, metacognition

This paper tackles a core ITS challenge: predicting when students will disengage so tutors can intervene before it's too late. It introduces **engagement forecasting** as a supervised prediction task …

2026-05-20 · learning-analytics, engagement-metrics, k-12, efficacy-study, benchmark

Teachers virtually never test AI tutoring bots before student deployment; PromptDecipher enforces QA as a first-class activity by letting teachers edit bot responses directly. PromptDecipher addresses…

2026-05-19 · llm, teacher-role, faculty-development, edtech-platform, ai-literacy

LLM tutors achieve near-ceiling on correct steps but systematically over-reject valid-suboptimal reasoning and over-validate incorrect solutions — precisely where adaptive tutoring matters most. This …

2026-05-19 · llm, generative-ai, benchmark, scaffolding, feedback-loop

Closed-loop ITS with multimodal affective scoring (facial, vocal, textual, oculomotor) produced significant presentation skill gains (Cohen's d = 0.39-0.90, N=204) over 30 days. This paper presents on…

2026-05-19 · affective-computing, multimodal, higher-ed, professional-training, efficacy-study

Socratic questioning, reflection prompts, misconception checks, and mandatory pauses produce better K-12 engagement than directive answer-giving AI tutors. SocratiCode demonstrates a participatory des…

2026-05-19 · llm, generative-ai, k-12, scaffolding, student-experience

Codify (also referred to as "AI Tutor") is a web-based [[intelligent-tutoring]] platform for programming education that integrates conversational AI, adaptive assessment, and learning analytics. It le…

2026-05-17 · generative-ai, higher-ed, llm, programming-its, scaffolding

This paper exposes a critical failure mode in using LLMs as simulated students for [[intelligent-tutoring]] development and evaluation. The authors introduce **misconception faithfulness** — the prope…

2026-05-16 · llm, generative-ai, benchmark, hallucination-risk, student-experience

KITE (Knowledge-Informed Tutoring Engine) introduces a [[intelligent-tutoring]] architecture that grounds its responses in course materials through a multimodal [[scaffolding|RAG pipeline]]. Unlike ge…

2026-05-16 · llm, generative-ai, scaffolding, formative-assessment, higher-ed

> Are Agents Ready to Teach? A Multi-Stage Benchmark for Real-World Teaching Workflows **Chen et al. (2026)** — Multiple institutions. Under review.…

2026-05-15 · agentic-ai, benchmark, generative-ai, llm, scaffolding

> Sycophancy is an Educational Safety Risk: Why LLM Tutors Need Sycophancy Benchmarks **Kasneci & Kasneci (2026)** — Position paper. arXiv cs.AI/cs.HC.…

2026-05-15 · hallucination-risk, llm, generative-ai, benchmark, over-reliance

> Evaluating Prompt Injection Defenses for Educational LLM Tutors: Security-Usability-Latency Trade-offs **Maiorano (2026)** — arXiv cs.CR/cs.AI.…

2026-05-15 · llm, generative-ai, regulation, hallucination-risk, student-experience

In a large-scale quasi-experiment with 635 students (grades 5-8), hybrid human-AI tutoring produced substantial gains over AI-only tutoring: +25% time on task, +36% skill proficiency, and +61% standar…

2026-05-14 · k-12, personalized-learning, efficacy-study, learning-gains, equity-in-ai-education

This paper bridges LLM-based dialogue tutoring and interpretable student modeling. By mapping opaque LLM representations to **Item Response Theory** parameters — student ability (θ) and question diffi…

2026-05-13 · knowledge-tracing, llm, personalized-learning, feedback-loop

This paper introduces **PLKT** (Probabilistic Logical Knowledge Tracing), which replaces deterministic vector embeddings with **beta-distributed probabilistic embeddings**, allowing explicit represent…

2026-05-13 · knowledge-tracing, learning-analytics, student-experience

This paper addresses a critical but under-examined issue in KT systems: **selection bias** from non-random exercise recommendations. Prior KT methods train on observed logs using standard empirical ri…

2026-05-13 · knowledge-tracing, learning-analytics, personalized-learning, student-experience

**Cognitive Agent Compilation (CAC)** is a framework that uses a strong teacher LLM to compile problem-solving knowledge into an explicit, inspectable target agent. Unlike end-to-end LLM tutoring appr…

2026-05-11 · llm, adaptive-learning, personalized-learning, pedagogical-llm-training, adaptive-learning-systems

> **Authors:** Kirk Vanacore, Ryan S. Baker, Avery H. Closser, Jeremy Roschelle **Year:** 2026 **Venue:** arXiv (cs.HC) > Synthesizes intelligent tutoring systems research and generative AI into a kee…

2026-05-11 · generative-ai, llm, scaffolding, personalized-learning, feedback-loop

> **Authors:** Youjie Chen, Xixi Shi, Xinyu Liu, Shuaiguo Wang, Tracy Xiao Liu, Dragan Gašević **Year:** 2026 **Venue:** arXiv (cs.CY) > Large-scale analysis (N=11,406 students, 200 classes, 10 instit…

2026-05-11 · generative-ai, student-experience, engagement-metrics, higher-ed, learning-analytics

Most AI-based educational tools adopt a one-on-one tutoring paradigm, pairing a single LLM with a single learner. Yet decades of learning science — from Vygotsky's Zone of Proximal Development to Band…

2026-05-11 · llm, scaffolding, student-experience, collaborative-ai-tutoring, agentic-ai-ecosystems-higher-education

> A framework for evaluating AI tutoring systems that extends beyond pedagogical quality of feedback to measure what students actually *do* with that feedback — whether they act on it and whether they…

2026-05-09 · efficacy-study, higher-ed, benchmark, engagement-metrics, llm

> Stub — pending source ingestion. Lifelong learning and AI support for continuous education beyond formal schooling.…

2026-05-09 · lifelong-learning, personalized-learning, professional-training, llm, higher-ed

> A bibliometric–systematic review of AI tools in STEM education: > Systematic review (2005–2025) mapping how AI tools scaffold and co-regulate metacognitive development in STEM classrooms through bib…

2026-05-08 · metacognition, stem-education, learning-analytics, posthumanist, ai-literacy

> ECNUClaw is an open-source framework by Zhou, Li & Zhang (2026) for building **learner-profiled intelligent study companions** in K-12 education. The system maintains a **five-dimension learner prof…

2026-05-08 · k-12, personalized-learning, llm, student-experience, self-regulated-learning

> Key limitations exist in both LLM-based tutoring and conventional Deep Knowledge Tracing (DKT): > Combining neural networks with symbolic educational knowledge for interpretable, data-efficient, and…

2026-05-08 · learning-analytics, adaptive-learning, pedagogical-safety, mastery-learning, pedagogical-llm-training

> As reinforcement learning personalizes instruction in intelligent tutoring systems, there is no formal framework for pedagogical safety — a critical gap. > First formal framework for defining and de…

2026-05-08 · pedagogical-safety, adaptive-learning, reinforcement-learning, adaptive-learning-systems, metacognition

> **SCRIPT** (Deriyeva, Dannath, Paassen, 2026) implements an intelligent tutoring system for **Python programming** in a German university context, filling a gap in prior ITS which rarely supported P…

2026-05-08 · stem-education, higher-ed, adaptive-learning, formative-assessment, student-misconceptions-conditionals-loops-taxonomy

> **From Prototype to Classroom** (Elhaimeur & Chrisochoides, 2026) describes a tutoring system for quantum computing that bridges the gap between dense mathematical formalism and limited qualified in…

2026-05-08 · adaptive-learning, higher-ed, knowledge-tracing, stem-education, adaptive-learning-systems

> **MuTSE** (Roscan et al., 2026) addresses a critical need in **Intelligent Tutoring Systems (ITS)**: delivering content at the right reading level for each learner. > Human-in-the-loop evaluation fr…

2026-05-08 · nlp-education, adaptive-learning, human-in-the-loop, generative-ai, human-in-the-loop-ai

Automated question generation leverages NLP and LLMs to create educational assessments at scale. Wei & Stamper (2025) introduced the **generate-then-validate** paradigm, reducing hallucination by 62% …

2026-05-08 · nlp-education, assessment, llm, stem-education, automated-grading

Culturally Relevant Pedagogy (CRP), introduced by Gloria Ladson-Billings (1995), centers marginalized students' cultural references in curriculum design. Wang et al. (2025) demonstrate that **LLMs can…

2026-05-08 · equity, curriculum-design, culturally-sustaining-pedagogy, k-12, ai-literacy

Equity in AI Education addresses systemic disparities in access to, representation within, and benefits from AI educational tools. Three critical dimensions emerge: Wang et al. (2025) found that **78%…

2026-05-08 · accessible-learning, culturally-sustaining-pedagogy, equity, ai-literacy, higher-ed

Teacher AI competency encompasses the knowledge, skills, and dispositions required for effective AI integration in educational contexts. Emerging frameworks identify three competency dimensions: Zhang…

2026-05-08 · ai-literacy, faculty-development, k-12, higher-ed, generative-ai

> A design framework for educational AI systems structured around four agentic paradigms: **reflection**, **planning**, **tool use**, and **multi-agent collaboration**. Proposed by Kamalov et al. (202…

2026-05-07 · agentic-ai, ai-education, benchmark, pedagogical-llm-training, human-in-the-loop-ai

> Zerkouk, Mihoubi & Chikhaoui (2025) systematically analyzed qualified studies from 2010–2025 across: > A comprehensive systematic review of AI-based Intelligent Tutoring Systems (2010–2025) reveals …

2026-05-07 · benchmark, efficacy-study, higher-ed, k-12, pedagogical-llm-training

> Conventional LLM safety benchmarks focus on toxic outputs, jailbreaks, and bias. In education, the primary risks are quieter: > "Solving problems correctly and avoiding toxic language does not make …

2026-05-07 · pedagogical-safety, adaptive-learning, k-12, higher-ed, llm

> ProPACT constructs a real-time model of pair collaboration using three signals: > Most adaptive learning systems are individual-centric and reactive. **ProPACT** treats **collaboration itself as the…

2026-05-07 · adaptive-learning, scaffolding, higher-ed, formative-assessment, learning-analytics

> Two critical gaps in dialogue-based Knowledge Tracing (KT): > Most LLM-based dialogue tutoring systems produce opaque predictions. Huang et al. map raw LLM logits into **student ability (θ)** and **…

2026-05-07 · adaptive-learning, personalized-learning, learning-analytics, k-12, llm

> Current AI tutoring systems treat each session as independent. Adaptive systems use real-time knowledge tracing (e.g., [[knowledge-tracing-irt|IRT-based models]]) but rarely retain a longitudinal st…

2026-05-07 · llm, personalized-learning, adaptive-learning, generative-ai, lifelong-learning

> Ostrowska, Kukla & Majstrak (2026) present an AI tutoring system **integrated into the Moodle LMS** designed to scaffold students from surface-level fact recall to deep conceptual understanding thro…

2026-05-07 · higher-ed, edtech-platform, llm, scaffolding, adaptive-learning-systems

> When LLMs process STEM problems that require interpreting diagrams, graphs, or schematics alongside text, their accuracy degrades substantially. This effect is: > General-purpose LLMs achieve near-c…

2026-05-07 · llm, stem-education, higher-ed, k-12, equity

> 1. **Desirable difficulties** — General-purpose AI removes productive struggle; tutoring tools preserve it via graduated hints. 2. **Germane load** — Effective learning requires processing that feel…

2026-05-07 · llm, generative-ai, personalized-learning, scaffolding, adaptive-learning

> Integrating emotional awareness into AI tutoring systems can yield measurable pedagogical gains, but the same affective sophistication risks amplifying harms if learner agency is eroded by empatheti…

2026-05-07 · affective-computing, adaptive-learning, scaffolding, k-12, higher-ed

> Domain-specialized optimization can transform a mid-sized open-source model (Qwen3-32B) into a pedagogical domain expert that outperforms far larger proprietary systems — but only when training rewa…

2026-05-07 · llm, adaptive-learning, benchmark, ai-education, higher-ed

Tailoring educational experiences to individual learner profiles, including prior knowledge, learning pace, preferences, and affective states. AI enables personalization at scale, though the gap betwe…

2026-05-07 · personalized-learning, adaptive-learning, ai-education, llm, higher-ed

> Socratic dialogue — asking structured questions rather than providing answers — is one of the strongest pedagogical scaffolds for deep learning. When automated via AI, it produces measurable reasoni…

2026-05-07 · scaffolding, active-learning, higher-ed, stem-education, formative-assessment

> **Synthesis:** Raza & Farooq (2025) conduct a comprehensive content analysis of AI in education from 2020-2025, examining 100+ peer-reviewed articles through a three-layer framework: the genome laye…

2025-10-31 · ai-education, systematic-review, personalized-learning, generative-ai, learning-analytics