Tue.21.Apr.2026. Author: Peter Liljenberg
As AI solutions move from pilot to procurement, one question keeps returning: how do you show that an AI system can be trusted without slowing innovation down? A new initiative is trying to answer exactly that.
Currently developed within the EU’s Testing and Experimentation Facilities (TEFs), it will soon offer AI developers a credible, independent assessment label.
While it is not an official EU-certification, it will provide structured, third-party evaluations that companies can use in sales, marketing and procurement dialogues. In short, a signal of quality ahead of the longer-term processes of formal approval.
From principles to practice
The label will be built on several dimensions of trustworthy AI, ranging from fairness and transparency to robustness, accountability and risk management. These are rooted in EU ethics guidelines but translated into practical testing criteria.
Not every system is assessed against everything. The approach is modular: companies receive badges only for the dimensions that are relevant and tested. This keeps the process both rigorous and realistic, avoiding unnecessary burden while maintaining comparability.
“We don’t accelerate AI by skipping trust. We accelerate it by making trust clear, visible and relevant for those who need to use it,”says Morten Koed Rasmussen, project manager at Gate 21, Denmark.
A catalogue of evidence
Assessments will be linked to an emerging AI assessment catalogue, where methods, tools and results are documented. This creates transparency and traceability. A badge is not just a symbol, but an entry point to evidence.
Verification follows the same logic. Rather than pass or fail thresholds, the label shows that testing has been conducted and expert feedback provided. Fixed compliance lines are avoided in favour of contextual judgement.
A positive shift in trust
The ambition is simple but significant: to make tested AI visible and usable. For startups, the label becomes a selling point. For cities and buyers, it offers reassurance. And for Europe, it may help accelerate the uptake of responsible AI without waiting for formal certification systems to catch up.
In a field moving at speed, the AI assessment initiative suggests a different model of trust. Not absolute, not final, but documented, shared and good enough to move forward.
Co-funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union. Neither the European Union nor the granting authority can be held responsible for them.