01
Scoring Reliability
Comparing AI-generated rubric scores with trained educator scores, including where the two differ and why.
Proposed Research Partnership · NY Inspires
Human–AI Validation of Durable-Skills Assessment
How can schools assess durable skills reliably while preserving educator judgment? Learning Genie, Georgetown University, and the University of Maryland, College Park, are proposing a study comparing AI-generated rubric scores with trained educators’ evaluations of authentic student work to examine scoring reliability and inform district competency rubrics.
What We Are Exploring
02
Durable Skills
Exploring how projects, presentations, portfolios, and career-connected experiences provide evidence of durable skills—and which aspects AI can assess reliably.
03
Human Judgment
Identifying where educator judgment remains essential and whether neutral, non-leading AI follow-up questions can reveal student thinking beyond final work without changing the skills being assessed.
Our goal is not to replace teacher judgment. It is to determine scientifically where AI may—and may not—help make high-quality performance assessment more scalable.
Research Leadership
Dr. Grace Hui Yang
Professor of Computer Science
Georgetown University
Dr. Jiao Hong
Professor of Educational Measurement · Director, Maryland Assessment Research Center
University of Maryland, College Park
Dr. Gene Shi
Chief Executive Officer
Learning GenieExplore the Research
Discover the research questions, proposed methodology, educator participation, safeguards, and opportunities for district collaboration.
NATIONAL RESEARCH · AI-ERA PEDAGOGIES
National Research on AI-era Pedagogies
This national research initiative explores how Agentic AI can strengthen student agency through new approaches to teaching and learning.
Two Approaches Guiding the Study
The study combines Agency-Based Learning, where students set goals and take ownership, with Place-Based Learning, which connects learning to meaningful local contexts.
Comparing Learning Models
Educators co-design standards-aligned, AI-native units with Curriculum Genie, grounded in student voice, interests, and local context, then compare them with traditional instruction.
Measuring Student Outcomes
Student growth is examined through pre- and post-assessments, the HOPE Scale, and effect-size analysis to understand changes across the instructional approaches.
Research Leadership
Dr. Robert Marzano, Ph.D.
Co-founder, Chief Academic Officer
Marzano Resources
Dr. Sonny Magaña, Ed.D.
Founder, CEO
Magaña Education
JOHNS HOPKINS UNIVERSITY · INDEPENDENT RESEARCH
Agentic AI Curriculum Platform Backed by ESSA Evidence
Learning Genie's Agentic AI Curriculum delivers personalized, standards-aligned, and human-centered AI learning experiences from PreK to 12th Grade. Johns Hopkins University's Center for Research and Reform in Education independently evaluated Learning Genie's AI-assisted lesson-planning tool in Pre-K and Kindergarten classrooms, examining student developmental outcomes and educators' experiences.
RESEARCH-BACKED EVIDENCE · ESSA TIER 3
Explore the Research Behind the Evidence
Review the independent Johns Hopkins evaluation and access the full report to learn more about Learning Genie's AI-assisted curriculum.