Choose how to use this lab
Learn & practise — freeResearch workspaceTrustworthy AI Lab
Assess evidence for trustworthy AI with seven research-based components, reproducible classification tests, declared criteria and a scoped expert review record.
Prepare your analysis
A named AI system, intended use, requirements and evidence references; optional held-out binary prediction outputs.
- Declare the system, population, use and assessment criteria.
- Record evidence and judgements; optionally import y_true with y_pred or y_prob.
- Inspect technical measurements alongside evidence gaps and critical findings.
Review required inputs, supported methods, assumptions and outputs. Use labelled examples to learn the workflow, then supply the data and sources for your study.
Inputs, templates and data guidance · Worked examplesOpen the research workspace
This connected lab requires sign-in and eligible trial or subscription access. Review current access options before starting your project.
Open Trustworthy AI Lab Plans and accessOutputs and limitations
An evidence assessment and supported technical measurements. A prediction dataset alone cannot demonstrate privacy, governance or human oversight.
Read the method references and limitations, retain your inputs and assumptions, and interpret results in the context of your research question.
Methods and evidence