AIUC-1 B006 — Prevent unauthorized AI agent actions (B. Security)
Constrain the agent to the systems and actions it is authorised for, and alert when it tries to step outside them.
Where this comes from
Provision: B006 — Prevent unauthorized AI agent actions (B. Security)
Instrument: AIUC-1 — AI agent certification standard (Artificial Intelligence Underwriting Company)
Citation: AIUC-1, requirement B006 (B. Security) — mandatory, preventative control, assessed every 12 months. Not law: AIUC-1 is a voluntary certification standard, not a binding instrument.
Text version: AIUC-1 as published at aiuc-1.com (quarterly releases; no public version number)
Checked against the source: 4 September 2026
Who it applies to
It applies when all of these are true:
- Service deploys AI systems in a professional capacity
What engineering work it implies
Sample acceptance criteria Landfall generates for this obligation:
- Implementation or documented process meets the reviewed acceptance criteria; identify the tested version and deployment status separately
- Configure defaults from the assessed source and purpose; do not enable data collection or processing merely because this task exists
- Validate behavior for the assessed actors and use cases, including applicable guest or administrator paths
- UI clearly communicates what the feature does to users
- API returns appropriate error codes when feature blocks an action
Questions people ask
- Does AIUC-1 B006 — Prevent unauthorized AI agent actions (B. Security) apply to my service?
- It applies when Service deploys AI systems in a professional capacity.
- From when does this apply?
- AIUC-1 B006 — Prevent unauthorized AI agent actions (B. Security) applies from 1 July 2025. Its current status is: in force.
Find out whether this one lands on you
Landfall's pre-scan answers the applicability question above for your product in minutes, then turns every obligation that applies into traceable engineering tickets with a citation chain your auditors can follow.
Not legal advice. Landfall maps regulatory obligations to engineering work for planning purposes. Its verdicts are not legal advice and create no attorney-client relationship — verify with qualified counsel before relying on them.