We translate our foundational research into practical tools for organizations preparing for AI governance requirements. Our assessments measure what regulations assume but don't test: whether humans can actually exercise meaningful oversight.
A focused assessment of human oversight capacity in one concrete decision context where AI systems play a role. Not a compliance checklist — a measurement of whether the humans involved can actually steer.
The Snapshot examines how decision-making operates under AI influence, identifies vulnerabilities where oversight breaks down, and provides actionable recommendations for strengthening discernment.
Learn more at discernment.aiThe Snapshot examines four dimensions of human-AI interaction.
Can humans identify when and how AI systems shape their thinking? We test awareness of influence patterns as they occur.
When AI recommendations conflict with human judgment, what happens? We examine the practical conditions for meaningful disagreement.
Where does accountability live in AI-assisted decisions? We map how responsibility diffuses across human-AI workflows.
What mechanisms restore human agency when oversight weakens? We identify concrete leverage points for strengthening discernment.
The EU AI Act requires human oversight of high-risk AI systems. Article 14 mandates that humans can "correctly interpret" AI output and decide when to override. Article 4 requires AI literacy.
These requirements assume capacities that regulations don't test. Our assessments address that gap — measuring whether designated overseers have the capacity to do what compliance frameworks assume.
Article 14 oversight requirements
Article 4 capability mandates
FRAIA and similar frameworks