Human gates in Responsible AI: who carries the decision?

AI accelerates drafting, extraction, mapping, test creation and documentation. Accountability stays with humans: a named human reviews every regulatory interpretation, architecture decision, risk approval and production release.

The hard boundary

No agent makes autonomous legal, regulatory, AML/KYC, credit, underwriting, HR or medical decisions. It may retrieve information, draft, run checks and propose actions. Authorized humans review context, conflicts and evidence and carry the decision. High-risk or unclear cases go to a manual queue.

Practical gates

• Use-case classification before any access (prohibited/high-risk/transparency) with qualified legal.
• Evaluation with ground truth before every pilot; no release after a failed critical eval.
• Visible uncertainty, refusal and escalation as mandatory agent behaviour.
• Logged approvals with a named approver instead of silent automation.

These principles flow into every INFRA offer; the catalogue provides details.