AI Gate

AI systems compress decision-making. They remove the visible pause between intention and deployment, and diffuse responsibility across models, vendors, and teams. Decision Gate is mandatory before deploying any AI system whose decisions carry significant consequences.

Responsibility cannot be delegated to a model, a vendor, or a system.

Pre-deployment questions

Before deploying any AI system, every team must be able to answer all of the following with specificity. If any item cannot be answered clearly, deployment must not proceed.

#Question
01Why are we using AI here, and not a simpler approach?
02What decisions does this AI system make on behalf of users?
03Is the AI's decision visible and explainable to the person it affects?
04Can the outcome be corrected if the AI is wrong or biased?
05Who — by name — is responsible when this AI causes harm?
06Has the system been tested against adversarial and edge inputs?
07Is there a human override at every critical decision point?

Pre-deployment checklist

Critical — gate cannot pass without these
Critical
No named human decision owner

There must be a named individual who accepts accountability for every significant outcome this AI system produces. Not a team. Not a vendor. One person.

Critical
Decision outcome is non-reversible or non-compensable

If the AI makes a decision that cannot be corrected, compensated, or appealed, this risk must be explicitly evaluated and accepted at the gate.

Critical
Decision invisible to the person it affects

If the AI makes a decision about a person and that person cannot see or understand it, explicit ethical justification must be documented before deployment.

High — required before proceeding
High
No human override path exists

At every critical decision point, a human must be able to intervene, override, or stop the AI. This path must be tested, documented, and accessible.

High
No adversarial or bias testing completed

The system must be tested against edge cases, adversarial inputs, and demographic subgroups. Results must be documented before deployment.

High
Training data undocumented or unreviewed

Data provenance, known biases, and exclusions must be documented and accessible to the decision team.

Medium — address before or during deployment
Medium
Scope creep governance not established

If the AI is likely to expand beyond its original scope, governance must be established before deployment — not after expansion has already occurred.

Apply it today — no setup required.

No tools · No certification · No infrastructure