Doug Holton

RE: @alfiekohn@sciences.social

The Long Push To Blame Systemic Problems On Individuals
https://www.sciencefriday.com/articles/its-on-you-book-excerpt/

Equitable Group Work in Undergraduate Biology Courses: Leveraging a Complex Instruction Framework to Identify Pedagogical Strategies
https://www.lifescied.org/doi/10.1187/cbe.25-07-0154
#Teaching #EdDev #STEMeducation

Free #Ebook: ABC Learning Design: Active, blended, connected and beyond
https://uclpress.co.uk/book/abc-learning-design/
More info & resources: https://abc-ld.org/
#LearningDesign #OpenAccess #EdDev

The LLM Fallacy: Misattribution in AI-Assisted Cognitive Workflows
https://arxiv.org/abs/2604.14807
"a cognitive attribution error in which individuals misinterpret LLM-assisted outputs as evidence of their own independent competence, producing a systematic divergence between perceived and actual capability"
#AIEd #psy #hci #LLM

The DOJ has delayed its accessibility deadline (April 24) by one year (2027): https://www.federalregister.gov/documents/2026/04/20/2026-07663/extension-of-compliance-dates-for-nondiscrimination-on-the-basis-of-disability-accessibility-of-web
In other news, the free PAVE tool for fixing PDF accessibility has been upgraded to 2.0. It can still be very time-consuming and complex to fix a single PDF, however. You have to fix all the regions, fix the reading order, etc. of every single page: https://pave-pdf.org/
Open weight LLM AI PDF OCR models are improving: https://huggingface.co/spaces?category=ocr&sort=trending
#accessibility #a11y

Academic Integrity in the Age of AI https://www.cambridge.org/core/elements/universitypress-integrity-in-the-age-of-ai/8652D952D1C480A46996183626BE3DD7 is free until April 20th. It's about 60 pages long, summarized in the NotebookLM infographic below. Hopefully it's not totally giving up on #OnlineLearning as the infographic suggests.
I did a presentation on a similar topic a few months ago: Strategies for Reducing Student Misuse of AI https://docs.google.com/presentation/d/1htjhjS7-cLx8BfdL2aZZ40B8opUz1ckZedcxmeJYUco/edit?usp=sharing To me, the main underlying key is student motivation (slides 18-22)
#AIEd #AcademicIntegrity #Teaching

SafeTutors: Benchmarking Pedagogical Safety in AI Tutoring Systems
https://arxiv.org/abs/2603.17373
"the primary risk is not toxic content but the quiet erosion of learning through answer over-disclosure, misconception reinforcement, and the abdication of scaffolding"
"We uncover that all models show broad harm; scale doesn't reliably help; and multi-turn dialogue worsens behavior, with pedagogical failures rising from 17.7% to 77.8%."
#AIEd #EdTech

EduQwen: Application-Driven Pedagogical Knowledge Optimization of Open-Source LLMs via Reinforcement Learning and Supervised Fine-Tuning
https://arxiv.org/abs/2604.06385
A fine-tuned open #LLM beats even Gemini on a #pedagogy benchmark. Unfortunately it doesn't appear to be released yet.
#AIEd

ISD-Agent-Bench: A Comprehensive Benchmark for Evaluating LLM-based Instructional Design Agents
https://arxiv.org/abs/2602.10620
Code & data: https://github.com/codingchild2424/isd-agent-benchmark
"benchmark comprising 25,795 scenarios that combines 51 contextual variables across 5 categories with 33 ISD sub-steps derived from the ADDIE model."

w/same author: Pedagogy-R1: Pedagogical Large Reasoning Model and Well-balanced Educational Benchmark https://dl.acm.org/doi/10.1145/3746252.3761133
#AIEd #LearningDesign #AIevaluation #EdTech

Examples of using #AIagent systems in education:
* Evolution of AI in Education: Agentic Workflows https://arxiv.org/abs/2504.20082v2
* DeepTutor https://github.com/HKUDS/DeepTutor
* OpenMAIC multi agent classroom https://github.com/THU-MAIC/OpenMAIC
* liacript teaching agent https://github.com/LiaScript/teaching-agent/tree/dev
* Claude education skills
https://github.com/GarethManning/claude-education-skills
https://github.com/narosemena/master-instructional-design
* Claw-ED https://github.com/SirhanMacx/Claw-ED
* CODE-GEN mc question generator https://arxiv.org/abs/2604.03926
#AIEd #AIeducation #EdTech

Knowledge graphs are useful representations for knowledge bases, #pkm, #AImemory systems, #GraphRAG, intelligent tutoring systems, etc., and usually implemented in graph databases. LadybugDB, a fork of the discontinued Kuzu, is a lightweight embedded (like SQLite) graph database: https://github.com/LadybugDB/ladybug
Sample applications in development: https://github.com/inventivepotter/dotmd & https://github.com/tejzpr/Smriti-MCP
See also Grafeo: https://github.com/GrafeoDB/grafeo
#AIEd #AIEngineering #KnowledgeGraph #GraphDB #graphdatabase

"AI petting zoo" resources for instructors: https://liascript.github.io/course/?https://raw.githubusercontent.com/edtechdev/aipettingzoo/master/README.md#1
#EdTech #AIEd #EdDev