Breadcrumb
- Home
- AI Hub For Education
- Research Study Repository
- School Leader
Search and Filter
The Repository
School Leader
Research synthesis is AI-generated, human reviewed. Updated 11/2026.
Displaying 1 - 30 of 35
"We'Ll Fix It Later": Education, AI, And The Deferral Of Privacy In EdTech
Meghna Manoj Nair, Rachel Greenstadt. (09/2026). arXiv.
One-third of audited platforms made no meaningful AI disclosure despite visible AI features, and 73% provided only generic accountability and breach-response language.
Human-LLM Collaborative Inductive Coding For Conceptualizing K-12 Educator AI Use
Alex Liu, Min Sun, Lief Esbenshade, Michael Xiao, Victor Tian, Zachary Zhang, Kevin He. (07/2026). arXiv.
Collaborative inductive coding produced a final hierarchical codebook of 72 items across 19 categories and six domains, after human coders identified five missing codes.
A Principled Way To Think About AI In Education: Guidance For Educators And Policy Makers On Action Based On Goals, Models Of Human Learning, And Use Of Technologies
Noah D. Finkelstein. (07/2026). arXiv.
Educational goals must explicitly prioritize developing individual capacities and societal infrastructure over mere workforce preparation when deploying artificial intelligence tools.
From Chaos To Clarity: A Framework For Program-Level AI Learning Outcomes
Grace Barkhuff, Ian Pruitt, William Gregory Johnson, Rodrigo Borela, Ben Rydal Shapiro, Anu G. Bourgeois. (07/2026). arXiv.
Program-Level AI Learning Outcomes (PLAI-LOs) create a structured alignment connecting institutional priorities, program outcomes, course goals, and individual assignment policies.
Validating AI-Generated Classroom Observations: Reliability, Accuracy, And Limits Of LLM-Based Pedagogical Judgment
Carolina Melo, Javiera de la Maza, Matías Recabarren. (06/2026). Computers and Education: Artificial Intelligence.
Repeated evaluations showed substantial random variability, and no model achieved uniformly high reliability across all instructional elements.
AI Diffusion Gaps: Unequal Integration of AI Across K-12 School
Christopher Campos, John D. Singleton. (06/2026). NBER.
A higher share of disadvantaged students is associated with lower AI integration, scoring 0.07 to 0.11 standard deviations lower on an integration index per standard deviation increase in disadvantage.
Practitioner Beliefs And Behaviors In AI-Enhanced Education: Dot Framework Survey Evidence
David C Gibson, Mary Elizabeth Azukas, Gerald Knezek. (05/2026). arXiv.
Factor analysis revealed three core belief dimensions: AI functional capabilities, oversight and governance, and instructor collaboration and planning, explaining 54.4% of total variance.
Prediction Model Of Motivators And Demotivators Of Integrating Large Language Models In Software Engineering Education: An Empirical Study
Maryam Khan, Muhammad Azeem Akbar, Jussi Kasurinen, Estefanía Martín-Barroso. (05/2026). arXiv.
Programming and debugging assistance, personalized learning, and AI as a learning partner were rated as the strongest motivators for adopting large language models.
Governing generative AI in higher education: a global Delphi study on policy and practice
Helen Crompton, Diane Burke, Christine Nickel, Aras Bozkurt, Fengchun Miao, Mike Sharples, Jeffrey Alan Greene, David Parsons, Lucy Gill-Simmen, Adam Edmett, Mark Pegrum, Inge de Waard, Curtis Jay Bonk, Manuel B. Garcia, John H. Curry, LeeAnn Lindsey, Mohan Yang, Stephen Marshall, Maha Bali, Nellie Deutsch, Suzaan le Roux, Mourad Benali, Mohd Ali Bin Samsudin, Hasan Tinmaz, Matthew L. Bernacki, Mari van Wyk, Lenandlar Singh, Agnes Chigona, Lance Eaton, Junhong Xiao, Johanna Velander, Jinhee Kim, Francisco Bellas, R. Rajalakshmi, Andréia de Bem Machado, Agnieszka Palalas, Sean Yu. (05/2026). International Journal of Educational Technology in Higher Education.
A global consensus framework identified eight core areas for higher education generative AI policies, including academic integrity, ethics, privacy, access, literacy, integration, oversight, and institutional support.
Facet: Multi-Agent AI Supporting Teachers In Scaling Differentiated Learning For Diverse Students
Jana Gonnermann-Muller, Jennifer Haase, Nicolas Leins, Moritz Igel, Konstantin Fackeldey, Sebastian Pokutta. (05/2026). arXiv.
Surveyed teachers gave favorable ratings to didactic structure and clarity across differentiated worksheets, scoring highest for profiles targeting low-motivation learners.
Mapping AI Programs In The U.S: A Status Report From Early 2026 And An Analysis Of AI Majors And Minors
Felix Muzny, Carolyn Jones, Carter Ithier, Hasnain Sikora, Hrutika Harshadbhai Patel, Carla E. Brodley. (05/2026). arXiv.
Of 563 surveyed U.S. institutions, 44% offered at least one AI program, with concentrations being the most common format at 32.7% of all programs.
Agentic AI Ecosystems In Higher Education: A Perspective On AI Agents To Emerging Inclusive, Agentic Multi-Agent AI Framework For Learning, Teaching And Institutional Intelligence
Vidya K Sudarshan, Anushka Sisodia, Reshma A Ramachandra, Batra Sia, Josephine Chong Leng Leng. (05/2026). arXiv.
Approximately 52% of reviewed studies focused on task-specific AI tools and 28% used single-agent systems, while multi-agent systems (12%) and agentic multi-agent systems (8%) were rarely explored.
Knowledge without Wisdom: Measuring Misalignment between LLMs and Intended Impact
Michael Hardy, Yunsung Kim. (04/2026). arXiv.
Language models agreed more strongly with one another than with human experts, but their ratings often negatively aligned with actual student learning gains.
Machine Learning Grade Prediction Using Students' Grades And Demographics
Mwayi SONKHANANI, Symon CHIBAYA, Clement Nyirenda. (02/2026). arXiv.
Logistic regression and random forest classifiers achieved test accuracies of up to 94.6% to 96% for predicting student pass/fail outcomes.
Trustworthiness Of Legal Considerations For The Use Of Llms In Education
Sara Alaswad, Tatiana Kalganova, Wasan Awad. (02/2026). arXiv.
The United Kingdom and European Union demonstrate the most comprehensive regulatory frameworks, explicitly addressing all six core trustworthiness dimensions.
Interoperability In Ai Safety Governance: Ethics, Regulations, And Standards
Yik Chan Chin, David A Raho, Hag-Min Kim, Chunli Bi, James Ong, Jingbo Huang, Serge Stinckwich. (01/2026). arXiv.
AI governance in education across all four countries relies primarily on data privacy and child protection rules rather than mature, dedicated technical AI standards.
Bidirectional Human-Ai Alignment In Education For Trustworthy Learning Environments
Hua Shen, NYU Shanghai, New York University. (12/2025). arXiv.
Trustworthy educational AI requires a two-way alignment process where systems reflect shared human values while educators and learners develop competencies to critically evaluate and guide those tools.
An Exploration Of Higher Education Course Evaluation By Large Language Models
Bo Yuan, Jiazi Hu. (12/2025). arXiv.
A fine-tuned model (Llama-UKP) achieved strong correlation with human expert ratings (Pearson's r = 0.754, Spearman's rho = 0.843), outperforming a general-purpose model.
Principled Design Of Interpretable Automated Scoring For Large-Scale Educational Assessments
Yunsung Kim, Michael Hardy, Joseph Tey, Candace Thille, Chris Piech. (11/2025). arXiv.
The interpretable framework scored within 0.06 quadratic weighted kappa of state-of-the-art uninterpretable models across 10 assessment items and outperformed multiple automated scoring baselines.
Fairaied: Navigating Fairness, Bias, And Ethics In Educational Ai Applications
Zhipeng Yin, Sribala Vidyadhari Chinta, Zichong Wang, Matthew Gonzalez, Wenbin Zhang. (11/2025). arXiv.
Educational AI biases operate across individual, group, and multi-level dimensions, where interventions targeting one level can unintentionally create or amplify disparities at another.
Measuring Teaching With Llms
Michael Hardy. (10/2025). arXiv.
Three sentence-embedding models achieved human-level performance, with correlations exceeding 0.65 for general teaching practices and surpassing average human rater correlations for math instruction.
Ai-Generated Essays: Characteristics And Implications On Automated Scoring And Academic Integrity
Yang Zhong, Jiangang Hao, Michael Fauss, Chen Li, Yuan Wang. (10/2025). arXiv.
The e-rater automated scoring engine systematically assigned higher scores to AI-generated essays than human raters did across all evaluated models.
Quantum Annealing For Staff Scheduling In Educational Environments
Alessia Ciacco, Francesca Guerriero, Eneko Osaba. (10/2025). arXiv.
On the real-world school instance, the hybrid quantum annealing solver reached the optimal schedule found by the exact solver across all five runs in about 15.16 seconds.
Geolog-Ia: Sistema Conversacional Sobre Tesis AcadéMicas
Micaela Fuel Pozo, Andrea Guatumillo Saltos, Yeseña Tipan Llumiquinga, Kelly Lascano Aguirre, Marilyn Castillo Jara, Christian Mejía-Escobar. (10/2025). arXiv.
Geolog-IA achieved an average BLEU score of 0.87 when comparing its generated answers to actual database query results.
A Principled Way To Think About Al In Education: Guidance For Action Based On Goals, Models Of Human Learning, And Use Of Technologies.
Noah D. Finkelstein. (10/2025). arXiv.
Educational practices with AI must center on three core human objectives: cultivating discernment, fostering empathy for collaboration, and building a learner's sense of self and agency.
Securing Educational LLMs: A Generalised Taxonomy of Attacks on LLMs and DREAD Risk Assessment
Farzana Zahid, Anjalika Sewwandi, Lee Brandon, Vimal Kumar, Roopak Sinha. (08/2025). arXiv.
Token smuggling, adversarial prompt injection, direct prompt injection, and multi-step jailbreaks present the most critical cybersecurity risks to educational large language models.
Inclusive Education with AI: Supporting Special Needs and Tackling Language Barriers
Ricardo Fitas. (04/2025). arXiv.
AI translation and captioning tools allow multilingual learners to follow instruction and communicate, helping them perform closer to native-speaking peers on comprehension assessments.
Where Assessment Validation and Responsible AI Meet
Jill Burstein, Geoffrey T. LaFlair. (11/2024). arXiv.
Responsible AI practices can be directly mapped to classical validity inferences, such as using system prompts as item specifications to support domain description.
Exploring Generative AI Policies in Higher Education: A Comparative Perspective from China, Japan, Mongolia, and the USA
Qin Xie, Ming Li, Ariunaa Enkhtur. (07/2024). arXiv.
All four countries show positive attitudes toward generative AI, but Japan and the United States emphasize human-centered approaches while China and Mongolia prioritize national security.
Comprehensive AI Assessment Framework: Enhancing Educational Evaluation With Ethical Ai Integration
Selcuk Kilinc. (06/2024). arXiv.
School bans on generative AI are easily circumvented, and AI text detection tools show poor accuracy that drops to 22% against adversarial techniques.
