Services
Algorithmic Audit
A comprehensive assessment of risks associated with a specific use case of an algorithmic system. We audit systems of all kinds, including generative AI, automated decision systems, predictive models, and facial recognition.
Uses our Ethical Matrix framework
Identifies high-priority issues around fairness and performance
Delivers recommendations for measuring and mitigating risks
Outputs / Deliverables: Algorithmic Audit Report
Pilot: Quantitative testing for regulatory compliance and more
Analysis to measure bias in algorithmic systems, using data from live deployments or test data.
Proprietary, patent-pending cloud based analysis platform
Incorporates inference methodologies, so we can measure gender and race/ethnicity bias even if you do not possess this data
Double-firewall privacy protection: we never see personally-identifying information; client never sees individual inferences
Outputs / Deliverables: Bias Audit Report (e.g. for NYC Local Law 144); custom analysis reports
AI Governance + Risk Management Consultation
We help build core infrastructure to help you use AI responsibly.
Develop an organization-level approach to measuring and managing the risks of deployed AI systems, especially with regard to bias/discrimination
Review and recommend governance policies, processes, and structures
Assistance with procurement of AI technologies from vendors and other 3rd parties
Outputs / Deliverables: Documentation of risk management structures; diligence reports on prospective vendors
Cockpit Design
Running a deployed AI system is like piloting a plane: to fly safely, you need critical information in real time. To make an effective cockpit, you must understand what can go wrong in flight, and include dials/gauges that monitor for those risks. We help build cockpits for your AI systems.
Identify and/or construct metrics that address key risks
Calibrate thresholds and develop mitigation tactics
Can work with existing analytics platforms or assist with your custom implementation
Education + Training
We teach people how to think critically about AI and algorithmic systems, how to identify blind spots, and about the craft of auditing.
Workshops
Bespoke training for audit and risk management teams