LABS · IN DEVELOPMENT

Open Cultural Evidence Framework.

A shared way to move from objects, archives, and documents to assertions, evidence, uncertainty, contradictions, and human review.

/labs/open-frameworkCultureOS Labs · 2026

THE QUESTION

What should responsible AI scholarship have to show?

Its sources, contradictions, uncertainty, machine interventions, review history, and final accountable authority.

WHAT IS OPEN

A common evidence language.

01

Evidence & Assertion Model

Claims, sources, contradictions, confidence, provenance, and review.

02

Human Review Protocol

Where automated research stops and scholarly authority begins.

03

AI Research Audit Standard

How model-generated research is documented.

04

Cultural AI Evaluation Toolkit

Measures for accuracy, uncertainty, labor, and failure cases.

05

Reference implementation & guides

A small deployable example and implementation guidance for institutions.

THE BOUNDARY

Open the protocol. Commercialize the execution.

Evidence schemas, review methods, and evaluation are open. Multi-agent orchestration, entity resolution at scale, production interfaces, hosting, and security are built for institutions. We say so plainly.

PERMANENT SCHOLARLY ARCHIVE

Archived at CERN / Zenodo · DOI: 10.5281/zenodo.22828367

The complete specification and evidence schema are permanently archived on the European Organization for Nuclear Research (CERN) Zenodo repository for open scholarly citation across universities, museums, and research libraries.

Suggested Citation: Culture OS Labs (2026). Culture OS: Open Cultural Evidence Framework. Version 1.0. DOI: 10.5281/zenodo.22828367 → https://doi.org/10.5281/zenodo.22828367

THE FIELD WE BUILD ON

We are not the first to try any single piece of this.

These projects solve important parts of the problem. We build on them; our contribution is the integration. Acknowledgments, not partnerships.

START HERE

Working on the same problem?

Book a first conversation. It costs nothing.