CinderpointAISAFEMACHINE › BASE
SAFEMACHINE Module 3  ·  Readiness Assessment

BASE

By Waydell D. Carvalho  ·  Cinderpoint  ·  First published April 2026
Definition
BASE is the readiness check organizations run before deploying AI governance. The acronym names the four prerequisites: Budget, Authority, Skills, Expectations. Without all four in place, SAFEARC and CARG produce paperwork rather than oversight. BASE answers a single question: is this organization actually equipped to govern this system?

Why BASE comes first

Governance frameworks fail for two distinct reasons. The first is bad framework design, the rules do not match the problem. The second, and more common, is that a perfectly sound framework gets handed to an organization that cannot execute it. The team has no funded time, no decision authority, no relevant expertise, and no aligned expectations from leadership. The result looks like governance on paper and behaves like nothing in practice.

BASE catches that failure mode before SAFEARC and CARG run. It is a four-part readiness check. If an organization cannot pass BASE, the right answer is not to skip it and hope. The right answer is to fix the gaps first.

The four prerequisites

B . Budget
Is governance work funded as work?

Governance that runs on volunteer overtime is not governance. Budget covers staff time on a named cost line, the tools and platforms required, and the cost of the reviews and audits the framework calls for. If governance has no line in the budget, it has no priority in practice.

A . Authority
Can the governance function actually stop a deployment?

A review board with no veto is a recommendation board. Authority means the named role or committee can hold a deployment, require a fix, or escalate to leadership without going through the team being reviewed. Without authority, the framework produces opinions that get overridden whenever speed is the priority.

S . Skills
Do the people doing the review understand the system?

Governance requires people who can read a model card, evaluate a fairness audit, interpret a security report, and understand the regulatory context. Skills can be hired, trained, or retained as outside expertise. They cannot be skipped. A review by people who do not understand what they are looking at is a signature, not a review.

E . Expectations
Has leadership agreed in advance that governance can say no?

Expectations are the organizational agreement that governance is allowed to slow things down. If executive sponsors expect a rubber stamp, every framework on earth turns into one. Expectations get set explicitly, in writing, before the first review, with a documented escalation path for disagreements.

How BASE plugs into SAFEMACHINE

BASE is the gate before SAFEARC. The SAFEMACHINE platform asks the BASE questions first, scores the four prerequisites, and either clears the organization to begin SAFEARC reviews or returns a readiness gap report. The gap report names what is missing and what needs to change before the rest of the platform produces useful output.

The point is honesty about what governance requires. A SAFEARC review run by a team with no budget, no authority, no skills, and no leadership backing produces a document that looks like governance and functions like decoration. BASE refuses to pretend otherwise.

About the author
Waydell D. Carvalho

Founder of Cinderpoint Systems LLC. M.S. Artificial Intelligence (MSAI), M.S. Management (MSM). Researches how systems fail under speed, opacity, and scale.

More by this author SSRN ↗ Zenodo ↗