Free AI review practice kit for adult educators and workplace trainers

The Human Approval Gate is a free toolkit that helps educators, trainers and teams decide when a person must review AI-generated work before it has real-world effects. It was developed by AI trainer Keith Staggers. The kit gives practical guidance for the moment before an AI-assisted output is sent, published, used to change a record or used to commit money.

What is a human approval gate?

The kit defines a human approval gate as a controlled decision point. At this point, a named, qualified person checks defined evidence against written criteria before an AI-assisted output may affect real work. A gate is recommended when AI output could:

• be sent or published outside a private draft space;
• alter a record, schedule, permission, account or system;
• affect a person’s access, opportunities, care, employment, finances or reputation;
• create a charge or financial commitment;
• involve sensitive information or professional guidance; or
• trigger another action that is hard to reverse.

The CLEAR test

At the core of the method is the CLEAR test, a five-point check for a sound review step:

• Consequence held: no real-world action happens before the gate.
• Listed reviewer: a named, qualified person owns the decision.
• Evidence checked: the reviewer inspects defined evidence, not just a polished result.
• Allowed outcomes: accepting, revising, rejecting and escalating all remain possible.
• Record and recovery: the decision, the exact version, the reason and a safe response are documented.

What the kit includes

The English materials include an eight-minute video lesson, a nine-page practitioner guide and a one-page printable worksheet for defining a gate and recording decisions. Ten practice cases cover situations such as missing information, conflicting details, unsupported claims, tool outages and changes made after approval.

A German edition is designed for adult education and continuing training. It includes a six-page guide, an editable worksheet, facilitator notes for a 30-minute session and ten practice cases with answers. A workshop announcement serves as the worked example. No chatbot account is needed to use it.

Access and limitations

All materials can be downloaded free of charge, without registration, and are licensed under Creative Commons Attribution 4.0 (CC BY 4.0), so they can be adapted for training. Version 1.0.2 is publicly archived on GitHub and Zenodo (DOI: 10.5281/zenodo.22145298). The guide draws on public sources, including the NIST AI Risk Management Framework.

The author presents the Human Approval Gate and the CLEAR test as proposed practical methods. They are not validated research instruments, a standard, a certification or legal advice. Users should adapt them to the laws, privacy rules and organisational policies that apply to their own context.

Explore the Human Approval Gate toolkit.