Breadcrumb
Search and Filter
The Repository
Analyzing
Research synthesis is AI-generated, human reviewed. Updated 11/2026.
Displaying 1 - 30 of 317
When Evaluators Cry Wolf: Lessons From Production LLM-As-Judge Evaluation In Educational AI
Chris Rohlfs, Rodrigo Vergara Bosse, Priscilla Rain Hopper, Keanon O'Keefe, Patrick Russell. (09/2026). arXiv.
Across 21 deployed evaluators, the calibration strategy cut confirmed false-positive flags by 99% and raised per-flag precision from 0.6% to 48.9%.
Fairness Theatre: Evaluating Post-Hoc Fairness Interventions In Vendor-Controlled Early Warning Systems
Kelly McConvey, Angelina Zhai, Rebecca Li, Shion Guha. (09/2026). arXiv.
Post-hoc fairness interventions redistributed disparities without consistently reducing them, leaving error burdens for marginalized groups intact or worse while dashboard metrics appeared to improve.
Early Prediction Of AI-Assisted Cheating Risk In Online Exams Through Learning Analytics
Gokhan Akcapınar. (09/2026). arXiv.
Logistic regression predicted AI-assisted cheating risk from first-eight-week learning management system traces with 73.1% accuracy, a balanced accuracy of 0.718, and an area under the curve of 0.763.
Three Pathways Of Student-AI Interaction: Constraint-First Design For Higher-Order Thinking
Fatima Tuz Zahra. (09/2026). arXiv.
Undergraduate student interactions fell into three paths: 46% Passive Review, 18% Direct Question, and 36% Strategic Dialogue in primary classifications.
"Code Is Cheap. Show Me The Talk.": Lessons From Teaching And Managing AI Coding Tool Usage In A Visualization Course
Zhongzheng Xu, Taehyun Yang, Fumeng Yang. (09/2026). arXiv.
About half of students' prompts in AI-coding labs focused on refining existing charts, while generating new code accounted for roughly one third of turns.
Making The Invisible Visible: A Framework For Reflective AI Use In Software Engineering Education
Ali Shakiba, Thomas Chaffey. (09/2026). arXiv.
Retained students reported high perceived utility of AI assistance (mean 4.42/5) and high confidence in evaluating AI-generated outputs (mean 4.52/5).
Argus: Academic Integrity In The Era Of Generative AI
David Racovan, Ajay Rawat, Christopher K. May, Jeffrey A. Turkstra. (09/2026). arXiv.
In Spring 2026, 45.7% of enrolled students were flagged for suspected LLM-assisted code development following automated triage and manual review.
Characterizing LLM-Based Family Education Through The Lens Of Activity Theory: A Scoping Review Of The Hci Literature
Lan Luo, Yuqi Liang, Jie Cai, Anqi Wang, Dongyijie Primo Pan, Muzhi Zhou, Chun Yu, Pan Hui. (09/2026). arXiv.
Most reviewed studies focused on child–parent interaction (83%) and centered on language, narrative, AI literacy, and relational development rather than mathematical or scientific inquiry.
"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.
Towards Synergistic Teacher-AI Interactions With Generative Artificial Intelligence
Mutlu Cukurova, Wannapon Suraworachet, Qi Zhou, Sahan Bulathwela. (09/2026). arXiv.
Among reviewed generative AI studies supporting teachers, operational teaming was most common (62.5%), followed by transactional (17.5%), situational (17.5%), and praxical teaming (2.5%).
AI-Assisted Social Story Intervention For Special Education: The Design Of Adapted Stories
Buyankhishig Enkhjargal, Himanshi Lalwani, Hanan Salam. (09/2026). arXiv.
AI assistance suggests it can lower the burden of preparing Social Stories while supporting more individualized story creation.
Aware But Unprepared: Measuring The Security Awareness-Behavior Gap In Student Use Of LLM-Generated Code With Bifrost
Jaehwan Park, Seonhye Park, Kyungchan Lim, Doowon Kim, Doowon Kim. (09/2026). arXiv.
Over 95% of undergraduate students accepted insecure AI-generated code during programming tasks, despite 83.6% having completed security coursework.
AI-Assisted Assessment Of Experimental Physics Laboratory Reports: Potential, Limitations, And Support For Teaching Practice
Marcos Abreu, Cecilia Stari, Arturo C. Marti. (09/2026). arXiv.
Differences between AI and instructor assessments may stem from difficulties retrieving and interpreting equations, graphs, tables, and units rather than purely reasoning limitations.
AI-Written Admissions Essays Are Widespread But Penalized
Calvin Isley, Johann Gaebler, Sharad Goel. (09/2026). arXiv.
By 2025, 56% of applicants submitted at least one primarily AI-generated essay—reaching 69.3% among international applicants—despite an explicit prohibition against AI use.
Finding Common Mistakes In Modelling With Mathematical Formalisms Using LLMs
Lilian Killich, Marko Schmellenkamp, Fabian Vehlken, Thomas Zeume. (09/2026). arXiv.
On a propositional logic data set of 6,106 student pairs, the workflow discovered transformation clusters explaining 84.44% of student mistakes, including 15.03% not captured by previous hand-picked analyses.
Exploring K-12 Teachers' Perceptions Of Students' Relationships With AI Companions: Boundaries, Intervention Strategies, And Design Implications
Qing Xiao, Wenhan Xie, Ziyu Deng, Ruiwei Xiao, Ziyue Feng, Xie He, Shiyu Zhang, John Stamper, Hong Shen, Xinying Hou. (09/2026). arXiv.
Most teachers found AI tutors and study buddies acceptable, while intimate roles like romantic partners, friends, or parents raised concerns about eliminating developmental friction.
AI-Augmented Inquiry And Regulation In Hybrid Systems: A Control Allocation Architecture For Preserving Epistemic Agency In Hybrid Human-AI Cognition
Jochen Kuhn, Peter Gerjets, Ulrich Trautwein, Jeffrey A. Greene, Sarah Malone, Patrik Vogt, Tim Fütterer. (09/2026). arXiv.
Hybrid cognition can become destabilized through seven interacting mechanisms, including control redistribution, germane effort attenuation, calibration drift, delegation drift, representational passivity, dual-threshold instability, and motivational-affective drift.
Do Assessment Instruments Measure The Same Thing For Humans And LLMs? A Latent Structure Analysis
Alona Strugatski, Licol Zeinfeld, Giora Alexandron. (08/2026). arXiv.
Tests showed systematic differences between human and language model factor structures, indicating the exams may not measure the same constructs in humans and AI.
INSIDE the Student's Mind: Jointly Modeling Latent Reasoning and Action in LLM Student Simulators
Rose Niousha, Minwoo Kang, Narges Norouzi. (08/2026). arXiv.
Fine-tuning models on reconstructed internal dialogue achieved the closest match to real student code distributions and the highest reasoning alignment, reaching up to 57.9%.
Challenges For Musical Education In The Age Of AI And Digital Transformation
Jean-Pierre Briot. (08/2026). arXiv.
Generative AI can produce complete, stylistically coherent musical pieces from short text prompts, altering compositional practice alongside digital audio workstations.
From Precision Medicine To Precision Education: A Vision For AI-Powered Student Digital Twins, Preventive Student Success, And Career-Aligned Academic Pathways
and Kaushik Dutta. (08/2026). arXiv.
Predictive risk scores alone do not improve student outcomes without organizational capacity, rapid human outreach, and targeted financial supports.
A Bottom-Up Taxonomy Of Student Discourse With A Socratic AI Physics Tutor
Syed Furqan Abbas Hashmi, N. Sanjay Rebello. (08/2026). arXiv.
Student discourse was concentrated in a few areas, with the top 25 of 357 categories covering roughly half of all student messages.
Eduzone: A Framework For Evaluating LLM Safety For K-12 Students And Teachers
Junyeong Park, Jieun Han, Haneul Yoo, So-Yeon Ahn, Jinsung Yoon, Alice Oh. (08/2026). arXiv.
Models were more vulnerable to education-specific risks than conventional harms, with average attack success rates reaching 80.3% for student academic misconduct and 73.0% for excessive cognitive load.
Virtual Tutoring with Computer-Assisted Learning: An Experiment in Take-Up and Learning
Philip Oreopoulos, Ruochong Dong, Nina Low. (08/2026). NBER.
Offering the tutoring raised Khan Academy practice by about 10 minutes weekly and led to an average math assessment score gain of 0.055 to 0.06 standard deviations.
Can AI deliver appropriate support for diverse student profiles?: A large-scale evaluation
Sonsoles López-Pernas, Eduardo Araujo Oliveira, and Kamila Misiejuk, Fitsum Gizachew Deriba, Rogers Kaliisa, Mohammed Saqr. (08/2026). Computers in Human Behavior: Artifical Humans.
The evaluated language models showed limited sensitivity to student characteristics and learning analytics indicators when generating support recommendations.
The Theoretical Foundation Of Socratic Tests: Dynamic, Multimodal, Conversational Examinations
Ilya Mikhelson. (07/2026). arXiv.
Over 80 percent of surveyed students agreed that the conversational AI understood their responses and provided effective scaffolding.
Structured AI Demonstrations And Student LLM Use In Engineering Mechanics: Study Design And Preliminary Results
Shuang Geng, Helen Lallos-Harrell, Jiya Ashar, Thomas J. McKenna, Annwesa Dasgupta, Caleb Farny, Emma Lejeune. (07/2026). arXiv.
Undergraduate use of artificial intelligence for engineering mechanics rose from 37% weekly use at baseline to 51% by the end of the semester.
Agreement Is Not Quality: Blind Expert Verification Of Human And LLM Qualitative Coding When Human Consensus Is Not Ground Truth
Alex Liu, Lief Esbenshade, Michael Xiao, Victor Tian, Zachary Zhang, Kevin He, Min Sun. (07/2026). arXiv.
A blind expert preferred human and model coding at statistically indistinguishable rates (51.5% versus 48.5%), despite models having lower agreement with human coders than humans had with each other.
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.
Generative AI Availability, Grades, And Student Satisfaction At A Large University
James M. Zumel Dumlao, Meng Wang, Zhonghan Xie, Junyao Hu, Ivan Bar, George Chaney III, Henry Gold, Misha Teplitskiy. (07/2026). arXiv.
Effects of ChatGPT's release on student grades were small and non-robust, ranging from 0.010 to 0.071 grade points depending on how pandemic trends were modeled.
