Abby (Rubinson) VollmerMember of the Board of Directors
Research institute · Research orgs · Est. 2023
Humane Intelligence is a U.S.-based 501(c)(3) nonprofit that breaks down barriers to AI deployment for social good by collaboratively designing and running rigorous, human-centered evaluations of AI systems. Its programs focus on (1) AI red teaming (semi-structured adversarial testing to find vulnerabilities and improve safety and effectiveness), (2) AI contextual evaluations (mixed-method, bespoke evaluations tied to a specified problem space), and (3) bias bounties (community-driven challenge design intended to examine and improve models, systems, and datasets with input from impacted communities).
The organization also operates a web application used for data collection in hosted red teaming and contextual evaluation services, and it is working to release its red teaming application as open source in 2026. Strategically, Humane Intelligence positions evaluation as a lever for “accountability, responsibility, and fairness” across high-stakes domains, and it invests in building repeatable evaluation practices (including taxonomy/ontology work) while convening multi-sector communities to operationalize evaluation methods in real deployments.
No indexed openings right now. Check the careers page ↗.
A look at key concepts of both and why both need funding to help ensure AI does no harm.
2026 SciPy Conference RecapHumane Intelligence's actionable insights on key conversations and takeaways at SciPy 2026.
AI Evaluations for Everyone: Bringing Human-Centered AI Evaluation to LinkedIn LearningRead about the LinkedIn Learning course from Humane Intelligence that explores how non-engineers can evaluate AI systems.
The Global Divide Over Data GovernanceWe explore three models of governance and how that applies to AI
Folie à Deux in the Machine: When Models Share HallucinationsRiddles leave models with nowhere to hide.
Bias Bounties on Zindi Preview: Geospatial Data and Climate RiskA blogpost about Humane Intelligence's Bias Bounty on Geospatial Data and Climate Risk
OSS AI red teaming app release updatesRead about OSS AI red teaming app updates, including our design, front-end and backend work, as well as how to get involved
Mental Health as a Hidden Cost of AI SystemsA blogpost about why mental health must be treated as a core consideration in AI design and evaluation
Our New Members of the Board of Directors and Advisory GroupHumane Intelligence welcomes new members of the Board of Directors and Advisory Group.
Tribal Innovation Summits Adversarial Stress Testing Exercise ReportIndigenized Adversarial Stress Testing FrameworkHumane Intelligence announced that the AI in Public Health working group is scoping an AI public health fund, with pitch materials described as going live in late August 2026 and updated as the fund progresses.
What is the relationship between AI safety and AI evaluations?Humane Intelligence published a thought leadership post explaining how AI evaluations are positioned as part of AI safety by definition, and describing differences between frontier AI safety research and evaluations of consumer-facing models and systems.
AI Evaluations for Everyone: Bringing Human-Centered AI Evaluation to LinkedIn LearningHumane Intelligence announced its LinkedIn Learning course introducing human-centered AI evaluation methods, including a release window described as July 1 through September 30, 2026.
We’re releasing our AI red teaming app under an OSS licenseHumane Intelligence announced that a proposal to release its AI red teaming app under an OSS license was fully funded with support from Google.org’s Scientific Ecosystem team.
Updates about our OSS AI red teaming app releaseHumane Intelligence posted updates indicating progress toward releasing an improved version of its AI red teaming app under an open source software (OSS) license, with planning described as previously announced.
Reflecting on the Accessibility Bias BountyHumane Intelligence reflected on its Global Accessibility Bias Bounty and reiterated funding/support described in relation to Google.org and Zoom Cares for expanding prizes.