The geopolitics of data standards: what is being decided in Europe and why it matters

When you click, the video loads from YouTube in privacy-enhanced mode. You can also watch it on YouTube.

Why data standards are geopolitics

Knut Blind, coordinator of StandICT_eu, opened the session with a simple idea: data standards are no longer a purely technical matter. They determine whether a digital single market for data can exist, whether Europe keeps its digital sovereignty, and whether there is enough reliable data to train and run artificial intelligence.

The EU framework: from more data to quality data for AI

Camila Kawa · DG CONNECT, European Commission

The 2020 European data strategy aimed to make more data available. With the Data Union Strategy the focus shifts: the goal now is to unlock high-quality data to train and run AI systems. The Commission is working on three fronts:

• Access to quality data: the problem is not just volume but quality and reliability (quality-assurance tools, synthetic data and standards).
• Regulatory simplification: the Digital Omnibus proposal would fold the essentials of the Data Governance Act, the Open Data Directive and the Free Flow of Non-Personal Data Regulation into the Data Act. It comes with model contractual terms and a Data Act legal helpdesk.
• International dimension: protecting the sovereignty of European data while connecting with ecosystems in like-minded countries.

The common European data spaces are the bridge between the data strategy and the AI strategy. Among the most mature, she cited the cancer imaging space, launched in 2023, which connects dozens of datasets and aims to exceed 60 million annotated images by 2027, along with the genomics and language spaces. Tourism, skills, the Green Deal, energy and agriculture come next.

Two standardisation requests are under way: a trusted data transactions framework supporting Article 33 of the Data Act, and another on data quality, with quantitative metrics such as completeness, accuracy and coverage.

Faster standards: revising Regulation 1025/2012

Carlos Rodríguez · DG CONNECT, European Commission

The Commission has proposed modernising the European standardisation system with three objectives: cutting the roughly six years a harmonised standard takes today, which is incompatible with the pace of AI; making it easier for SMEs to take part and giving legal certainty to access to harmonised standards; and strengthening Europe's presence in international bodies so that other powers do not set the course alone.

Trusted data transactions and the “data product”

Sebastian Steinbuss · Chair of CEN/CENELEC JTC 25 “Data”

This is the most practical part of the webinar. The Trusted Data Transactions standard has three parts: concepts and terminology (published), trustworthiness requirements (approved) and interoperability (in the final stage of becoming a harmonised standard). Trust rests on automatically reconciling the policies required with the verifiable claims each party presents.

The standard distinguishes who is accountable for a transaction from the software that acts on each party's behalf (see the explanation at 24:26):

• Accountable roles: the data rights holder, who decides how and to whom data is transferred; the data producer (a sensor, a device, a system); the data provider and the data user.
• Supporting roles: data intermediaries and governance authorities.
• Technical actors: each participant's agents, which manage metadata, negotiate the contract and carry out the transfer.
• Facilitating services: catalogue, discovery and trust services that issue and verify credentials.

The unit of exchange is not raw data but the data product: the data together with its version, provenance and lineage metadata, the terms of use (licences, permissions, obligations and prohibitions), the data model and the service level. It is managed across its whole lifecycle, from publication and negotiation to retirement.

It is complemented by a maturity model with six dimensions (governance, business, legal compliance, interoperability, data control and trust) and by the Vienna Agreement with ISO/IEC JTC 1/SC 38, so that these trust frameworks are approved in parallel as global standards.

Semantics and context-dependent quality: ETSI TC DATA

Diego López · Chair of ETSI TC DATA

In 2025 ETSI created a technical committee to bring its data work together, with three groups: information models and semantics, data infrastructure and sector applications (maritime transport, agriculture, 5G-Advanced and 6G). Its key building blocks are the SAREF ontology, so that data from different sectors can be understood across them, and the NGSI-LD protocol for managing and exchanging context information. A message worth keeping: data quality is not absolute; it depends on the use, which is why metrics verifiable at source are defined so each consumer can decide whether the data fits.

From open source to ISO: the IDSA route

Silvia Castelli · International Data Spaces Association

IDSA does not start from theory but from the more than 200 data spaces it tracks. Its cycle is: identify needs in real deployments, write open specifications in the Eclipse Data Spaces working group, test them with real components such as Eclipse Dataspace Components or Tractus-X, and take them to ISO through the PAS route. That is how the Dataspace Protocol (DSP), which lets connectors understand each other, and the Decentralized Claims Protocol (DCP), for identity and credentials, came about. Gaps remain: ontologies for monitoring compliance with usage policies, digital product passports and metadata that is interoperable across sectors and countries.

The minimum every data space must speak

Daniel Alonso · BDVA, Data Spaces Support Centre

The DSSC is entering a second phase focused on scaling up and real impact. After reviewing 130 standards, it recommends three common W3C pillars: ODRL to express usage policies, verifiable credentials for identity and DCAT to describe catalogues. He also announced the European Data Spaces Awards, assessed against the CEN/CENELEC maturity model.

Can Europe export its model?

In the closing panel, Knut Blind asked whether there will be a “Brussels effect” for data. Steinbuss answered yes on the technical side, because the protocols work the same everywhere, but not automatically on the organisational side, because each region understands sovereignty differently. López saw an opportunity to turn European principles into a seal of quality and trust, as happened with the GDPR. Castelli pointed out that the digital product passport will not work without agreeing standards with Asian manufacturers. Alonso stressed that cooperation with ISO and the W3C exists precisely so that Europe does not isolate itself.