Set the points where a person can intervene before a system output is acted on
Key facts
- Human involvement
- YellowClose supervision
- How AI is used
- From the decision path the lawyer supplies, list each step at which a person can pause, change or stop the system before its output is acted on. Map each step to guardrail 5 of the Voluntary AI Safety Standard, which calls for human control or intervention so there is meaningful human oversight across the life cycle. Do not decide which decisions must be stopped.
- What the lawyer must review
- The lawyer decides which outputs must not be acted on without a person, who that person is, and what they must see. Guardrail 5 is voluntary guidance, not a statute.
- Written for / version
- US
Open a jurisdiction version of this task to see which instruments apply: US
Who does what
| Role | What they do |
|---|---|
| Assistance | From the decision path the lawyer supplies, list each step at which a person can pause, change or stop the system before its output is acted on. Map each step to guardrail 5 of the Voluntary AI Safety Standard, which calls for human control or intervention so there is meaningful human oversight across the life cycle. Do not decide which decisions must be stopped. |
| Lawyer | Set the scope, supervise the work closely, verify every part relied on, resolve omissions and exceptions, and approve the exact output. |
Checklist
- Check each intervention point against the real decision path, not a generic lifecycle diagram.
- Confirm the map does not treat the Voluntary AI Safety Standard as a legal duty.
- Record who may intervene, what they see, and which outputs cannot proceed on the system alone.
Sources
- Department of Industry, Science and Resources: Voluntary AI Safety Standard — the 10 guardrails · 2026-10-22
These are suggested working checks. Confirm the required checking with the responsible lawyer before work starts; the checking approach does not establish permission to use AI.