The Collaboratory for Uncertainty Management in the Age of AI

What is CUMAAI?

The Collaboratory for Uncertainty Management in the Age of AI (CUMAAI, pronounced "koo-my") is an international research initiative originally founded by the Stanford University Center for Sustainable Development and Global Competitiveness (https://sdgc.stanford.edu), the IE University's Center for Entrepreneurship and Innovation  (https://www.ie.edu/entrepreneurship/), and researchers at the University of California, Berkeley (https://berkeley.edu/).

Drawing its name from the concept articulated by William A. Wulf in 1989 — "a center without walls, in which the nation's researchers can perform their research without regard to physical location" — CUMAAI operates as a living laboratory that crosses institutional, geographic, and disciplinary boundaries. It brings together expertise in innovation, entrepreneurship, cross-cultural learning, and generative AI to address one of the defining questions of our era:

The Central Research Question
As artificial intelligence becomes embedded in every domain of human activity, how do we preserve — and rigorously measure — human agency? 

Mission

CUMAAI's mission operates across three interconnected pillars:

The Agency Imperative: Why Measurement Matters

The most consequential challenge posed by generative AI is not technical — it is human. As AI systems increasingly perform cognitive tasks once reserved for people, the risk is not that machines will replace human workers. The risk is that humans will voluntarily relinquish the very capacities — reasoning, judgment, creative synthesis, emotional discernment — that define intelligent agency.

CUMAAI has placed human agency at the center of its research program. Specifically, the Collaboratory investigates:

This work has crystallized in the development of the Agency Capability Score (ACS) — a testable empirical construct that quantifies individual readiness for high-agency performance in AI-augmented environments.

The Role of PW Theory

PW Theory, developed through over a decade of doctoral (Esade) and post-doctoral research at IE University Business School and subsequently validated by researchers at Stanford and UC Berkeley, posits that an individual's comfort with uncertainty manifests distinctly across two life domains:

The intersection of P and W scores forms a unique "uncertainty signature" (e.g., P2W4, P3W1) that predicts how individuals make decisions, embrace innovation, delegate cognitive tasks to AI, and respond under conditions of deep uncertainty.

Critically, PW Theory functions simultaneously as a behavioral framework and as a measurement architecture. This dual function is what positions it as the empirical foundation for the Collaboratory's agency research:

Validated across 27 countries and more than 20,000 professionals, PW Theory is not only one of the most internationally tested frameworks in uncertainty science — it is the empirical backbone of CUMAAI's measurement program.

The Measurement Architecture: Quantifying Human Agency

A defining feature of the Collaboratory's research is its commitment to measurement. While much discourse about AI and human capability remains descriptive or anecdotal, CUMAAI is building a rigorous, replicable, and internationally scalable measurement infrastructure. This infrastructure is composed of four integrated layers:

Layer 1 — PW Theory Psychometrics (Uncertainty Profiling)

The CZone Scale provides the primary psychometric instrument for capturing individual uncertainty signatures. The instrument has been validated across cultures, industries, and career stages, yielding the W-score that feeds directly into ACS calculations.

Layer 2 — Agency Capability Score (ACS)

The ACS aggregates the W-score with validated measures of other factors to produce a composite agency readiness index. The ACS is currently introduced as a testable hypothesis in the Collaboratory's working paper "The AI Founder Paradox", with empirical validation underway through a University Lab research architecture.

Layer 3 — AI Behavioral Diagnostics (Cognitive Offloading Mapping)

The Collaboratory is developing diagnostic tools that detect patterns of AI-dependent versus AI-augmented behavior. These instruments map cognitive offloading tendencies onto PW Theory profiles, enabling researchers to identify at-risk populations and design targeted interventions before agency erosion becomes entrenched.

Layer 4 — Longitudinal & Cross-Cultural Validation

Through partnerships with a soon to be announced KSA University, CUMAAI is building a longitudinal dataset designed to track agency outcomes over time and across cultural contexts. This dataset will provide the evidentiary base for policy recommendations, organizational interventions, and academic publications.

AI and Cognitive Offloading

The rise of generative AI has created a profound paradox: the tools designed to empower human thinking can, if adopted unreflectively, systematically erode it. Cognitive offloading — the process of delegating mental effort to external technology — is not inherently problematic. The Collaboratory's research, however, identifies the conditions under which it becomes agency-diminishing:

CUMAAI addresses this phenomenon through the integration of PW Theory diagnostics with behavioral AI-use data, enabling the Collaboratory to:

This research strand directly informs the Collaboratory's flagship outreach initiative and is the empirical engine behind its emerging AI literacy measurement tools. See also Professor Paris de L'etraz's Keynote titled "Beyond Cognitive Offloading: Building Super Workers in the Age of AI" at the ICAN Ignite Symposium in Riyadh, Saudi Arabia on 27 January 2026.

"Thinking Humans": Global Outreach Initiative

To translate research into global impact, CUMAAI conducts an international outreach program: "Thinking Humans: Human Engagement in the Age of AI." This initiative engages students, educators, policymakers, and professionals in a movement for rational, conscious, and high-agency AI use.

Selected Research & Publications

CUMAAI's empirical outputs span peer-reviewed publications, conference proceedings, and working papers. Key works grounding the Collaboratory's current research agenda include:

Vision

The Collaboratory envisions a future in which humans evolve alongside AI — not beneath it. By embedding PW Theory and the Agency Capability Score into global education and workplace learning systems, and by building the measurement infrastructure required to track agency outcomes at scale, CUMAAI will help shape a generation that is:

AI-empowered, not AI-dependent
Comfortable with uncertainty, not paralyzed by it
Driven by curiosity and judgement, not by cognitive shortcuts
Measurably agentic - with evidence, not just aspiration

Through research, measurement, education, and global outreach, the Collaboratory will ensure that the future of intelligence — human and artificial alike — remains human at its core.

Institutional Research Partners

Join the Research

CUMAAI is actively building its global network of researchers, educators, partner institutions, and policy collaborators. If you are interested in contributing to the Collaboratory's measurement research, joining a collaborative session, or exploring institutional partnership, we welcome your engagement.