Implementation of the Trusted Data Observatory (TDO): A Global AI-Ready Metadata Platform for Trusted Data Discovery

Sponsoring Organization

Switzerland

Organization

Government

Supporting Organization(s)

Data Type

Official statistics Survey data Citizen data Private sector data Geospatial data Administrative data Artificial Intelligence Civil registration and vital statistics Census Monitoring and evaluation Data for SDG monitoring Cross cutting

Region

Global

Objective

The commitment aims to support the development of the Trusted Data Observatory (TDO), a global metadata platform designed to improve the discoverability, accessibility and reuse of trusted data and enhances AI-ready discovery of trusted data. The TDO will harmonise a Minimum Viable Metadata (MVM) set across national statistical offices and international organisations, improve discoverability for humans and machines, strengthen transparency, respect data sovereignty, and provide a neutral, global discovery layer that supports evidence-based policymaking and counters misinformation.

Description:

What problem is this commitment aiming to solve?

In an AI-enabled world, trusted official data often remains invisible to search engines and AI systems. Metadata is fragmented, inconsistently implemented, or not machine-readable, meaning authoritative data competes with misleading or inaccurate sources online. This weakens evidence-based policymaking and public trust. The core challenge is not the absence of trusted data, but the failure of digital systems to reliably discover and prioritise it.

Briefly describe the commitment including key activities, strategies and intended outcomes.

Switzerland has launched the implementation phase of the Trusted Data Observatory (TDO), following the 2024 feasibility study by the Federal Departement of Foreign Affairs and the Federal Statistical Office and the successful completion of Phase 1 and 2 - Full Report.

The TDO is a global metadata platform, not a data platform : all data remain with their original producers. It focuses on harmonised, standardised, open and machine-readable metadata aligned with MVM set to support AI-driven discovery of trusted data produced by National Statistical Offices (NSO) and UN-system organisations that follow the Fundamental Principles of Official Statistics (FPOS).

Through interoperability mechanisms, it will connect existing metadata catalogues of NSO and IO, enabling users and AI systems to identify who produces which datasets, their coverage, update frequency, and access conditions.

Describe how you will monitor progress and evaluate the success of this commitment?

2024 – Feasibility study completed

2025–2026

Agreement on Minimum Viable Metadata (MVM)
Governance framework and seed funding arrangements
Development and launch of Proof of Concept (PoC)
Prototype development and stakeholder onboarding
Financial Sustainability of the project

End 2026 – Q1 2027 – Review and refinement post-PoC

2027–2030 –

Broader onboarding of NSOs and IOs
Integration into post-2030 agenda discussions
Long-term governance and sustainability arrangements

Progress will be monitored through:

Adoption of the Minimum Viable Metadata (MVM) standard
Number and diversity of participating NSOs and IOs
Number of National Statistical Offices and International Organisations in the Proof of Concept
Prototype deployment and operational readiness
Evidence of improved discoverability of trusted data by AI systems

Periodic public reporting will accompany implementation phases.

Timeline

Following the initial conceptual and exploratory work, the TDO has now entered an institutionalisation phase. The current focus is on consolidating governance arrangements, strengthening the value proposition, and engaging strategic partners to support its development as a trusted international public good. This effort is being developed collaboratively with the Geneva data community and a growing network of international partners.

Contact Person

Benjamin Rothen international@bfs.admin.ch

Download