Artificial Intelligence in Education (AIED) · Springer, 2026
From Search Engine to Thought Partner: How Do Teachers Use AI Chatbots?
Extended analysis: What K–12 educators are actually prompting to AI — early findings (Stanford SCALE)
About
I am a Senior Researcher for the SCALE Initiative at Stanford University, affiliated with the GenAI in Education Hub and the National Student Support Accelerator.
My research applies natural language processing and causal inference to investigate effective practices in personalized instruction — developing tools that increase student engagement and learning.
Recent
News
-
Sep 2026
UpcomingConferenceSREE 2026 — Society for Research on Educational Effectiveness, Sep 23–26, 2026. One workshop and two papers accepted (more info to come).
-
Aug 2026
ReleasePreview release of TutorMoments, a benchmark for AI math tutoring, a collaboration between my team at SCALE Initiative and the Allen Institute for AI, which hosts the benchmark, and with the support of the Gates Foundation. More updates on this work coming soon!
-
Jun–Jul 2026
ConferenceFestival of Learning — Seoul, South Korea · Jun 27 – Jul 3, 2026
- Jun 28TalkWorkshop talk — Advancing the Science of Human and AI Tutoring Through Shared Infrastructure: A Collaborative Workshop.
- Jun 28TalkWorkshop talk — IRAISE 2026: Impactful and Responsible AI Systems for Education.
- Jun 30MentorshipServed as a mentor at the EDM PhD Consortium mentorship event.
- Jul 2PosterPresented “From Search Engine to Thought Partner: How Do Teachers Use AI Chatbots?” in the AIED Late-Breaking Results poster session.
Peer-reviewed
Publications
In progress
Working Papers
Preprint · Allen Institute for AI, 2026
When Help is Unhelpful: Evaluating AI Tutors for Productive Struggle
Introduces TutorMoments, a replay-based benchmark for evaluating whether AI math tutors scaffold when a student needs support and push for rigor when the student is ready. TutorMoments-Preview releases 462 de-identified tutoring transcripts and more than 1,500 teacher-annotated key moments. Supported by the Gates Foundation and Learning Commons.
EdWorkingPaper · 2026
Access Is Not Enough: Human Support Improves Engagement with AI Tutoring
EdWorkingPaper · 2024
Tutor CoPilot: A Human–AI Approach for Scaling Real-Time Expertise
EdWorkingPaper · 2025
How Tutors Share or Split Attention Across Students in Small-Group Tutoring
EdWorkingPaper · 2025
Understanding Disruptions to Virtual Learning: Causes of and Variation in Lost Instructional Time
Other works (drafts in preparation)Show fewer
Engagement RCT
Working paper · Draft coming soon
Enhancing Student Engagement in Virtual Math Tutoring Sessions: A Randomized Trial of a Tutoring Intervention
Draft coming soon
Test-taking behavior
Working paper · Job market paper
About Time: How Gender Differences in Test-Taking Behavior Explain the Gender Gap in Performance
Draft coming soon
Mindset & challenge
Working paper
Can a Growth Mindset Intervention Reduce the Gender Gap in Challenge-Seeking Behavior and Competitiveness?
Draft coming soon
Long-term outcomes
Working paper
Long-Term Outcomes of Affirmative Action: Evidence from a Selective Law School
Draft coming soon
Early-stage
Work in Progress
-
Moving Past Human Labels: Best Practices for Annotating Conversational Data
SCALE Initiative, Stanford University
-
Identifying Effective Tutoring Strategies Through NLP Analysis of High- and Low-Value-Added Tutors
SCALE Initiative, Stanford University
-
Practicing How to Teach with AI: Do Simulated Tutoring Sessions Improve Tutor Instructional Skills?
SCALE Initiative, Stanford University
In the press
Media Coverage
Access Is Not Enough: Human Support Improves Engagement with AI Tutoring has been featured in:
Tutor CoPilot: A Human–AI Approach for Scaling Real-Time Expertise has been featured in: