Trustworthy AI governance and audits for employment decision systems.
We help organizations evaluate and govern AI used in hiring, promotion, compensation, and performance management—so systems stay defensible, fair, safe, and operationally reliable.
What we study
AI Bias & Fairness
Assessing data, features, outputs, and decision workflows to identify bias risks and adverse-impact exposure across protected classes.
Validation Studies
Supporting job-relatedness, reliability, and defensible intended-use documentation for assessments and AI-driven models.
AI Governance & Policy
Designing policies, control owners, evidence requirements, and audit trails that hold up under regulatory scrutiny.
Threat Modeling & Safety
Mapping misuse cases, security threats, and failure modes—then defining controls and monitoring signals.
Selected work
NYC Local Law 144 Bias Audit Framework
Developed a replicable audit methodology for automated employment decision tools under New York City's landmark AI hiring law.
I-O Psychology x AI Governance Bridge Model
A novel framework mapping how industrial-organizational psychology fills critical gaps in current AI governance practice.
NAUTA Bias-to-Value Pipeline
A five-stage framework converting AI bias risk into measurable economic value through validated I-O methodology.
Recent research
Adverse Impact Analysis in AI-Driven Hiring: A Practical Framework
Working Paper · 2025
Construct Validity Gaps in AI Assessment Tools: Implications for Employment Decisions
Working Paper · 2025
The I-O Psychology Bridge: Closing Governance Gaps in Workforce AI Systems
Working Paper · 2024
AI governance intelligence
Why I-O psychology is the missing discipline in AI governance—and how it solves the bias problem that costs the U.S. economy billions. Data sourced from U.S. government agencies, peer-reviewed research, and federal policy frameworks.
The national security & economic problem
AI bias in employment decisions is not just an ethics concern—it is a national security vulnerability and an economic drag measured in billions.
Sources: DataRobot State of AI Bias Report (2024), MIT/NANDA research, PwC Global AI Study
Sources: White House E.O. 14110, EEOC, NIST AI RMF, Future of Privacy Forum, state legislatures
The I-O psychology solution: closing the governance gap
Most AI auditing treats bias as a technical problem. I-O psychologists bring the missing discipline—validated science of fair selection, job analysis, and human judgment—that transforms compliance checklists into defensible governance.
Novel Framework — I-O Psychology x AI Governance Bridge Model
This model maps how I-O psychology capabilities fill specific gaps in current AI governance practice. Each layer addresses a failure mode that pure technical auditing cannot solve.
bridges the gap
Model: Nauta Research Labs I-O x AI Governance Value Pipeline. Based on SIOP meta-analyses, EEOC enforcement data, and organizational performance research.
The NAUTA Bias-to-Value Pipeline™
A novel five-stage framework that converts AI bias risk into measurable economic value through I-O psychology methodology.
Detect
I-O MethodAdverse impact analysis, 4/5ths rule testing, subgroup disparity modeling across protected classes
EEOC Uniform Guidelines §60-3Diagnose
I-O MethodConstruct validation, job analysis alignment, criterion contamination/deficiency assessment
SIOP Principles · APA StandardsRedesign
I-O MethodEvidence-based selection redesign, alternative predictor evaluation, fairness-validity optimization
NIST AI RMF · Title VIIGovern
I-O MethodSOPs, control ownership, monitoring thresholds, escalation paths, and audit trail architecture
NIST GOVERN functionSustain
I-O MethodOrganizational change management, stakeholder adoption, continuous monitoring, and performance feedback loops
OD/Change ScienceOur approach
1) Define
- Decision scope and intended use
- Who is impacted and where risk concentrates
- Governance owner and audit requirements
2) Evaluate
- Data quality and integrity checks
- Bias/fairness and adverse-impact risk
- Reliability/validity evidence and limits
3) Reduce risk
- Redesign recommendations and workflow controls
- Human review gates where needed
- Monitoring signals and threshold alerts
4) Operationalize
- SOPs, training, and change management
- Documentation package and audit trail
- Ongoing review cadence and updates
Get in touch
For consulting inquiries, partnerships, or speaking engagements: