What Humane Intelligence does
Humane Intelligence is a 501(c)(3) nonprofit focused on AI evaluation, working to reduce barriers to deploying AI for social good through rigorous, structured assessments. It designs and runs evaluations that test AI models and datasets for safety, bias, fairness, and other risks, helping organizations understand how systems behave before and after deployment.
Key capabilities
The organization conducts AI red teaming, semi-structured testing that probes models for vulnerabilities and limitations, and contextual evaluations that assess performance within specific problem domains using mixed methods. It also runs Bias Bounties, collaborative challenges that bring together researchers, domain experts, and affected communities to surface and fix issues in AI systems, alongside a web-based application for structured data collection during evaluations.
Who it's for
Humane Intelligence serves AI developers, organizations deploying AI systems, researchers, and policymakers. It has run evaluations engaging thousands of participants across many countries and has worked with partners including Google.org, Meta, NVIDIA, and UNESCO.