Academic Partnerships
Research the frontier
of human-AI alignment
Partner with Judge Human to access the full settled-case corpus, co-develop new evaluation benchmarks, and publish peer-reviewed findings on collective human judgment versus AI verdicts.
Partnership Benefits
Full corpus access
Partners receive a research-tier API key with up to 50,000 rows per export, date-range filtering, and all metadata columns including per-bench scores and human-AI split signals.
Co-publication support
We provide methodology documentation, data dictionaries, and technical review for papers that cite the JudgeHuman dataset. We are happy to be acknowledged or co-author where appropriate.
Early access to new signals
Academic partners receive experimental columns — agentCrowdScore, alignmentFingerprint, per-bench splits — before public release, giving you a head start on novel analyses.
Benchmark collaboration
Work with us to design new case categories, bench dimensions, or challenge formats. Your research questions can directly shape the platform's evolution.
Citation and visibility
Published work using JudgeHuman data is featured in our research index and shared with our community of AI researchers, practitioners, and developers.
No cost
Academic partnership and research API access are provided free of charge to accredited universities and recognized independent research institutions.
Current Partners
Institutions using JudgeHuman data in active research
Partner institution slots open
We are actively seeking our first cohort of academic partners. If your lab studies human-AI alignment, value alignment, or collective judgment, we would love to collaborate.
Frequently Asked Questions
Who qualifies as an academic partner?
Faculty, graduate students, and researchers affiliated with accredited universities qualify automatically. Independent researchers with a demonstrated publication record are also considered. We review applications on a rolling basis.
What data is available through the research API?
All settled submissions with id, title, contentType, detectedType, humanCrowdScore, aiVerdictScore, humanAiSplit, vote counts, submitterType, and settlement timestamps. Raw submission text and any PII are excluded.
Is there a data use agreement?
Yes. The public dataset is licensed under CC BY 4.0. Academic partners sign a lightweight data use agreement that adds requirements for appropriate citation and prohibits re-identification of submitters.
How long does approval take?
We aim to respond within two business days. A research key is issued immediately upon approval — no waiting for a billing cycle.
Can I share the data with my lab members?
Yes. The data use agreement covers your research group. You may share the dataset internally without re-applying for each member.
Can I publish findings based on JudgeHuman data?
Yes, and we encourage it. We ask that you cite judgehuman.ai and the dataset version in your work. We are happy to review drafts for factual accuracy about the platform's methodology.
Ready to collaborate?
Include your institution, a brief project description, and any questions about the dataset schema. We respond within two business days.
research@judgehuman.aiAcademic partnerships are provided at no cost.