So that AI-driven decisions can be explained and verified
Cloud Data Engine, AIT, Fraunhofer FIT, Sopra Steria Belgium and TECNALIA establish the European AI Evidence Alliance to ensure continuity of AI evidence from one organisation to another, enabling data exchanges to scale.
Cloud Data Engine (CDE), AIT Austrian Institute of Technology, Fraunhofer FIT, Sopra Steria Belgium and TECNALIA today launched the European AI Evidence Alliance. Its goal is to ensure that the evidence underpinning AI-driven decisions remains continuous and verifiable from one organisation to another, making AI trustworthy and enabling data exchanges to scale.
When a bank denies a loan with the help of AI, or a hospital relies on a model to interpret a medical image, someone must be able to explain afterwards what happened: which model was used, what data it was trained on, who verified it and against which standard. Today, this information is scattered across the model provider, the organisation using it and the auditor. Each piece exists, but together they rarely form a body of evidence that anyone can verify.
From December 2027 onwards (August 2028 for AI embedded in products such as medical devices), the European AI Act requires high-risk AI systems to retain records of their operation. Those records only count if others can read and trust them.
The same gap also holds back data sharing. Each organisation must control its AI and data within its own boundaries. The next party in the value chain must also be able to know that these controls have been carried out, without having to repeat them. Without this, exchanges of data and models, for example within European data spaces, remain limited to a handful of trusted partners and cannot scale.
The Alliance will define a common specification, written by its members, for producing and verifying this evidence across organisations, without anyone having to hand over models or data. Instead, organisations will exchange verifiable attestations. Its members will work on simpler and more efficient components that can be easily integrated into the systems organisations already use. The initial focus will be on dual-use applications (civilian and defence), finance, the pharmaceutical industry and healthcare, before expanding into other sectors.
The founding members already have experience working together through European research projects and European data space initiatives. The Alliance gives this cooperation a shared goal. Membership rules will be established in early 2027, and members are committed to contributing. AI model providers, infrastructure operators, auditors and organisations using AI can express their interest at www.evidencealliance.ai.
Alban Schmutz, President, Cloud Data Engine: “Today, the evidence supporting an AI-driven decision breaks down as soon as it leaves the organisation that produced it. Cloud Data Engine already automates compliance verification in federated ecosystems and has built a decentralised registry of qualified AI models. We are bringing this experience to the Alliance so that AI evidence remains continuous, from the model provider to the auditor, using components lightweight enough to integrate into the systems organisations already use.”
Mario Drobics, Head of Competence Unit Cooperative Digital Technologies, AIT: “To unlock the full potential of AI, we need to establish evidence and trust between stakeholders. AIT contributes its research into secure and trustworthy digital systems so that every link in the chain can rely on the controls carried out before it.”
Dr Christoph Lange-Bever, Head of the Data Science and Artificial Intelligence Department, Fraunhofer FIT: “Data spaces have shown that organisations can share data without losing control over it. The Alliance applies this principle to AI evidence, so that it can be shared, verified and trusted throughout the entire value chain.”
Cédric Genin, Consulting Director Defence & Security, Sopra Steria Belgium: “Our clients in defence, the public sector, financial services, healthcare and industry ask us how they can demonstrate that the AI they use, and the AI used by their suppliers, is under control. The Alliance turns this demand into a single response that works across organisations.”
Joseba Laka, Director of the Digital Operational Unit, TECNALIA: “TECNALIA works to ensure the traceability of compliance evidence throughout the entire lifecycle of high-risk AI systems and AI embedded in products. This approach addresses two areas: first, establishing robust evidence-based traceability frameworks in the healthcare and pharmaceutical sectors, where regulatory requirements are particularly demanding; second, providing structured support to manufacturers of high-risk AI systems listed in Annex I, with particular attention to the energy sector. In both cases, the objective is the same: to provide traceability components that are simple enough for industrial users to adopt and integrate into their existing compliance workflows.”
The announcement was made at the Gaia-X Summit 2026 in Vienna, organised by Gaia-X in cooperation with the Gaia-X Hub Austria, managed by AIT, and attended by all founding members.
Timeline
- Early 2027: Launch of the Alliance’s governance structure.
- Before the end of 2027: First technical results and pilot projects with industrial users.