This evaluator supports Article 6(3) filter review after the Commission's 19 May 2026 draft high-risk classification guidance. It focuses only on material influence. It does not replace the full Article 6(3) Filter Decision Record.
Check whether the AI output steers the decision.
Use redacted facts. The tool logic runs in the browser, but do not enter sensitive, personal, privileged, or production data.
Material influence questions
Use this page to clarify one part of the Article 6(3) filter record.
Material influence means the AI output may shape a meaningful human decision through ranking, scoring, recommendation, exclusion, prioritisation, or strong flagging. This tool helps document that influence boundary; it does not decide legal classification.
No. Human review is not enough by itself. The practical test is whether the human can make an independent substantive assessment, override the AI output normally, document the override, and avoid rubber-stamping the model's recommendation.
Ask what the AI output does before the human decision. Check whether the output ranks, filters, flags, scores, or pre-selects people; whether reviewers see the output before deciding; and whether the vendor has evidence for the claimed boundary.
No. The Material Influence Evaluator is a supporting tool. Use it when the main Article 6(3) Filter Decision Record identifies material influence as the disputed issue or when a vendor claim depends on human review boundaries.
Retain the output type, affected decision, human visibility, override process, audit trail, reviewer role, vendor claim, and the final decision owner. Attach the record to the AI inventory and the Article 6(3) filter decision record.
Source basis: Regulation (EU) 2024/1689 and European Commission draft high-risk classification guidelines published on 19 May 2026. Last reviewed: 20 May 2026. This page is educational only and is not legal advice.