Resumes show what was, not what is now
Credentials rarely show whether someone can work effectively with AI today.
Assessment for AI-era work
Measure AI-ready
skills through
real work.
Simulation-based assessments that reveal judgment, communication, adaptability, and real-world AI collaboration.
Problem
Hiring methods miss the work itself.
Hiring teams need to see how candidates frame tasks, brief AI, verify output, adapt, and explain a decision under realistic constraints.
Interviews ask people to describe, not do
Words about work are not the work itself, and AI-generated materials hide the process.
Rigorous case assessment does not scale
Traditional work samples are expensive, inconsistent, and rarely capture AI-collaboration behavior.
Framework
Three skills shape AI-ready work.
The assessment separates practical AI use from the human capabilities that make AI-supported work trustworthy.
AI Fluency
Frame tasks, prompt AI, verify outputs, and explain when AI should or should not be used.
Relational Skills
Communicate, negotiate, and collaborate in AI-mediated work where human judgment still matters.
Adaptive Flexibility
Revise plans when goals, evidence, feedback, or tool behavior changes.
Design principles
A case-based assessment people can inspect.
The public story is not a black-box score. It is a visible chain from role-relevant simulation to evidence-linked review.
01
Simulation-Based Assessment
Candidates complete role-relevant simulations with realistic context, source materials, and deliverables.
02
Evidence Linked Evaluation
Scores connect to observable work: decisions, artifacts, rubric anchors, and reviewer notes.
03
AI Prepares, Human Decides
Agents structure evidence and surface patterns while trained reviewers own the final interpretation.
04
Review and Calibrate
Reviewers resolve disagreement, adjust anchors, and improve assessment quality over time.
Six practices
Six practices make AI-ready work visible.
The evidence map turns broad AI skill into behaviors a reviewer can see inside a realistic work sample.
Plan
Frame the problem, constraints, intended user, and where AI should or should not help.
Prompt
Give AI enough context, criteria, examples, and boundaries to support the work.
Probe
Check sources, assumptions, calculations, edge cases, and competing explanations.
Protect
Work within privacy, policy, fairness, and safety requirements.
Pivot
Revise the approach when evidence, feedback, or AI errors change the path.
Present
Explain the recommendation, tradeoffs, supporting evidence, and next steps.
Assessment evidence loop
AI prepares the evidence. Humans make the decision.
The system is designed around visible evidence, reviewer calibration, and partner learning rather than an opaque score.
Our team
Team
Rene Kizilcec
Associate Professor; Faculty Director, Future of Learning Lab
Rachel Slama
Associate Director, Future of Learning Lab
Philipp Kircher
Irving M. Ives Professor of Industrial and Labor Relations
Michèle Belot
Frances Perkins Professor of Industrial and Labor Relations and Economics
Thorsten Joachims
Jacob Gould Schurman Professor of Computer Science and Information Science
JR Keller
Associate Professor of Human Resource Studies
Steve Jackson
Professor of Information Science and Science & Technology Studies; Vice Provost for Academic Innovation
Paul Krause
Vice Provost for External Education; Executive Director, eCornell
Haocheng Zhang
Master's Student in Information Science
Diyang Lin
PhD Student
Jinsook Lee
PhD Student in Information Science