Evidence & public policy

Evidence into policy: briefs, standards & public positions.

AI policy decisions need more than technical claims or legal interpretation in isolation. TRAILAB connects research and evaluation evidence with the institutional questions, public values and practical constraints that shape policy choices.

Our policy work makes clear what the evidence supports, where uncertainty remains, and where recommendations depend on legal interpretation or normative judgement.

Evidence to decisions

A clear path from research to policy choices.

Frame the question

Define the decision, affected groups, institutional setting and the specific uncertainty the policy process needs to address.

Review the evidence

Bring together relevant research, evaluation findings and other sources, making the scope and strength of each source visible.

Interpret implications

Connect findings with legal obligations, fundamental rights, public value and implementation conditions while distinguishing evidence from interpretation.

Set out choices

Present options, trade-offs, limitations and open questions. Identify recommendations as reasoned judgements rather than direct experimental findings.

Policy contributions

Useful input for policy and regulatory work.

Policy briefs & evidence notes

Concise, research-grounded analysis for decision makers. A brief can set out the question, evidence base, competing interpretations, policy options, likely trade-offs, rights implications and remaining uncertainty. Unlike an evaluation report, it is organised around a public decision and its options rather than a system's test results.

Regulatory processes

TRAILAB can contribute technical and multidisciplinary evidence to public consultations, regulatory guidance, AI sandboxes and implementation discussions. The form of input depends on the process, question and available evidence.

Institutional positions

Evidence can inform positions on AI governance and public policy. Those positions should make their sources visible and distinguish empirical findings, legal analysis and normative commitments.

European AI standardisation

Connecting regulatory requirements with technical practice.

TRAILAB brings practical expertise shaped by active engagement in the CEN-CENELEC JTC 21 AI standardisation ecosystem. Standardisation provides shared technical language for requirements that need to be applied and assessed across different organisations and systems.

Our contribution connects AI Act requirements with evidence and practice in areas such as risk management, evaluation, data governance, transparency and human oversight. Standardisation work informs implementation; it does not replace formal conformity assessment.

Public positions & engagement

Make the evidence legible in public debate.

Briefings, position papers and public engagement can help institutions and wider audiences understand what research shows, what remains uncertain and which policy choices involve legal or ethical judgement.

European Parliament briefing · 2025

Technological Aspects of Generative AI in the Context of Copyright

A briefing on the relationship between protected training data, generative outputs and the technical tools that can contribute to copyright enforcement.

View the publication record

Position paper · 2026

Position: AI Usage Policies Should Be Aligned with International Human Rights Law

A position applying freedom-of-expression standards to assess whether provider rules on misinformation are clear, pursue legitimate aims, and are necessary and proportionate.

View the publication record

Transparency

Show the basis for each position.

Policy outputs should identify their evidence and authors, explain relevant limitations and uncertainty, and make clear when an argument moves from findings to interpretation or recommendation. Public positions also need safeguards for scientific independence and conflicts of interest.

Read about TRAILAB governance and independence

TRAILAB

Build evidence for public decisions.

Bring a policy question, regulatory process or institutional challenge to a multidisciplinary research team.