Consent rate benchmarks
What percentage of users share data, across which domains, in which integration contexts. Real numbers by data type and UX pattern. If you’re deciding whether to build a data-sharing flow, this is the data you need.
Every claim we make about data quality,
we measure first.
OpenDataLabs publishes research on AI training data quality, consented data collection, and the economics of human data licensing.
What we study
Contributor incentives and network sustainability
The economics of user data contribution at scale. Relevant for anyone building on a contributor network.
Portability and interoperability standards
How user data moves between systems without losing provenance. The standards question that AI teams, regulators, and platform operators will all have to answer.
Cross-domain enrichment effects
A contributor’s health data linked to their financial behaviour linked to their communication patterns tells you things none of those datasets tell you individually. We measure the performance uplift from multi-domain linkage and publish the methodology. That’s where ground truth data earns its price premium.
Data quality differentials: ground truth vs annotated vs synthetic
We measure the gap and publish the methodology. The difference matters most in health, finance, and language modelling. We quantify it.
Research
State of Context Report Q1 2026
April 14, 2026
Most AI deployments underperform because the missing piece is not the model but the context. This report breaks context into five measurable layers so teams can see where their AI systems are strong, exposed, and where to invest next. Read
Open Problems in AI Data Economics
October 30, 2025
We introduce data economics as a coherent field and define the open problems that haven't yet been formalised. Most AI economics research focuses on downstream effects. We argue you can't understand AI's economic trajectory without studying how data, compute, and labour interact at the production layer. Read
Model Influence Functions: Measuring Data Quality
August 1, 2024
A framework for attributing model outputs to training data contributions. The methodological foundation for pricing and valuing datasets in commercial AI pipelines. Read
Research collaborations.
We work with AI labs, universities, and enterprise data teams on applied research that needs ground truth data at scale. If your research requires that and you’re working on something that benefits the field, get in touch.