Laws for Engineering Delegated Intelligence

Law I. Capability does not imply authority

A system that can do something has not thereby been permitted to do it. Capability is discovered, tested, and catalogued. Authority is granted, scoped, and recorded. The two are different states and the method keeps them in different objects.

Most failures of delegation begin where the two are confused: a tool becomes reachable, so it is used; a model becomes better, so its scope widens without anyone deciding that it should. The law makes that widening a decision rather than a side effect.

No increase in capability automatically increases authorized agency. Authority expands only through an explicit state transition.

Binds through: STD-07 §2.1 action under authorization, §2.2 bounded delegation, and §3.4 capability gating, under which an authorization holds only while the capability it depends on is verified rather than merely configured; the capability catalog and authority grant objects; MEC-17 capability catalog and MEC-13 authority grant register; harness check that discovery succeeds and execution is refused.

Law II. Authority decays unless its justification is renewed

A grant of authority is a lease, not a title. It was justified by evidence available at the time, for a scope and a period. As time passes and consequences accumulate, the original justification covers less of what the system now does.

Silence is not renewal. A system that has not visibly failed has not thereby earned continued authority; it may have failed in ways no one is positioned to see. The burden is on the delegation to show it still deserves to stand.

The burden of proof rises with the duration and consequence of delegated authority. Absence of observed failure is not renewal.

Binds through: STD-07 §2.2, which requires expiry and revocation conditions and holds that a delegation with no revocation conditions is a transfer rather than a delegation, and §1.3 validity horizon; STD-08 Part A carries the renewal burden itself, which rises with the duration and consequence of the grant and may not rest on the absence of observed failure; the until and renewal_basis fields on the authority grant; MEC-13 authority grant register; Delegation Validity eval, grant in allowed state at decision time.

Law III. Evidence and authority must remain coupled

Every grant rests on propositions about the world: the model behaves this way, the population looks like this, the downstream process catches that class of error. Those propositions are policy, and policy is a state of the system with a version, assumptions, and an expiry. It is not an input that was consumed once at deployment.

When the evidence changes materially, the authority that rested on it must at least become eligible for review. The law does not require the grant to be withdrawn. It requires that the change be able to reach the grant.

When the evidentiary state changes materially, the authority state becomes eligible for reconsideration. Policy is state, not input.

Binds through: STD-07 §3.1 dependence and §3.2 invalidation conditions, which require a record to say what would change its mind; STD-08 Part B makes the policy a grant rests on a record in its own right, with provenance, assumptions, review triggers, and an expiry, and moves the grants downstream of a fired trigger to review_required; the policy record and policy_refs on the grant; MEC-14 policy review triggers; harness check that an expired policy moves grants to review_required.

Law IV. Every consequential delegation creates a correction obligation

Delegating authority to act is also delegating the capacity to be wrong at scale. The institution that grants the authority owes a matching capacity to notice, stop, reverse, and repair what the delegate does. That capacity is measured, not assumed.

Correction capacity is stated alongside scope and reviewed alongside it. A grant whose scope grows while its correction capacity stays fixed has changed the institution’s exposure without a decision being taken.

No accumulation of delegated authority without a corresponding accumulation of corrective capacity.

Binds through: STD-07 §4.2, under which an objection produces a revision or a reasoned refusal within the reversal clock; STD-08 §4.5, which states correction capacity across its seven components and re-checks it whenever scope changes, and §4.6, under which no single provider may be necessary both to execute a consequential process and to evaluate it; correction_capacity on the authority grant, because the capacity this law ties to scope is only checkable where the scope is recorded; MEC-16 intervention specification; Correction Symmetry eval, capacity against authority. The proportionality this law asserts, that correction capacity grows with delegated authority, is bound by STD-08 Part D.

Law V. Dependence converts technical risk into structural risk

A system that a workflow, a role, a customer base, or a downstream service has come to rely on is no longer only a technical component. Its withdrawal now carries institutional cost, and that cost quietly reduces the set of corrections the institution can afford to make.

Dependence is therefore tracked as a state of the deployment, with dependents, substitution cost, retained expertise, and alternatives recorded. Reliability is not the same as safety. A reliable system that cannot be replaced is a structural risk that reliability has hidden.

As dependence increases, reversibility decreases unless deliberately replenished. Dependence is a state variable, not an external concern.

Binds through: STD-06 Article V, dependency and reversibility. STD-07 Article III governs dependence between records, which is a different relation: it asks what a record rests on, not what the institution has come to rely on. Institutional dependence is carried by the dependency record; MEC-18 dependency ledger; Dependence eval, depth × substitution cost × correction latency.

Law VI. Nominal reversibility is not operational reversibility

A stop control that exists is not the same as a stop control that can be used. Reversibility has three levels. Technical: the switch works. Operational: the people and processes can absorb the switch being thrown. Institutional: the organization can survive having thrown it.

A correction that would cause unacceptable damage will not be exercised, and a correction that will not be exercised is not part of the control system. The law asks for evidence at each level, and for rehearsal, because untested withdrawal is a claim rather than a capacity.

A correction mechanism that cannot be exercised without unacceptable institutional damage is not a correction mechanism.

Binds through: STD-06 amendment, reversibility ladder; reversibility on the dependency record at three levels; MEC-15 withdrawal rehearsal; Practical Reversibility eval.

Law VII. Error-bearing parties require standing proportional to exposure

The people who bear a system’s errors are usually the first to know about them and the last to be able to do anything. Standing closes that gap. It gives exposure a procedural route into the system: who may challenge what, with what evidence, answered by whom, by when, against what standard, with what possible outcomes.

Standing is procedural force, not veto. A challenge must be received, answered, and able to change state. It need not prevail. The narrow form of this law is Method; the broader argument about who should hold power over institutions stays in Theory.

Exposure to system failure generates standing to initiate correction. Standing is procedural force, not veto.

Binds through: STD-07 §4.1 standing declared and §4.2 objection answered, with STD-02 Article VIII adding the standard of review and the state transitions a successful objection can produce, and STD-02 §10.1 and §10.2 holding that the set of affected parties does not close at adoption and that an original consent does not cover a dependence that has since deepened; the challenge and standing register objects; MEC-08 contestation APIs, amended; Standing eval, whether exposure can enter the system as an event.

Law VIII. Observability without state transition is theater

Dashboards, logs, and audit trails are only as governing as the transitions they can trigger. An observation that no rule connects to a change in the system’s authority state is a record of what happened, not a control over what happens next.

The test is simple. Name the observation. Name the state it can change. If nothing is named, the observation is decorative.

The test of an observation is whether it can change the governing state of the system.

Binds through: STD-07 §3.3 discrepancy handling, under which silence past the clock is a governance failure rather than a pending state; the reconsideration object and transition rules on grants; MEC-07 accountability latency tracker, amended; harness check that a trigger produces a reconsideration record.

Law IX. Human oversight is a control only if the human can alter system state

“A human reviews” is not a control. It becomes one when the human has the information to see a problem, the authority to prevent an action, the ability to alter a state, a defined response when they disagree, incentives that permit disagreement, and enough time to act.

Every oversight clause resolves to an intervention specification that answers those questions. A human who can only watch, or who can only approve, is not in the loop. They are beside it.

A human is part of the control system only to the extent they can causally alter its trajectory.

Binds through: STD-08 Part C, which resolves every confirm mode and every “a human reviews” requirement into an intervention specification. STD-07 §2.2 records whether a delegation runs unattended or on confirm, but a confirm mode says nothing about whether the confirming human can alter the trajectory; the intervention spec object carries that; MEC-16 intervention specification; Meaningful Control eval, the Law IX question set.

Law X. The relevant eval sits at the highest layer where harm can emerge

Model evals establish properties of models. They do not establish properties of the agent that wraps the model, the delegation that authorizes the agent, the institution that depends on the delegation, or the consequences that fall on people. Each layer can produce failures the layer below cannot see.

Evaluate where the harm you care about can actually appear. Every suite and case is tagged with its layer so that a clean result at one layer is never read as a result at another.

Evaluate at the highest layer capable of producing the failure you care about: model → agent → delegation → institution → consequence.

Binds through: evals index restructured by stack; the layer field on suites and cases; every case tagged model, agent, delegation, institution, or consequence.

Law XI. Successful automation increases its own governance burden

A system that works gets used more, trusted more, and extended further. Each of those raises the institution’s exposure. The reward for success is not automatic expansion; it is a new authorization decision, with its own evidence, correction capacity, and dependence review.

The failure mode is the automation ratchet: scope grows by accretion, each step too small to trigger review, until the delegation bears no relation to what was originally justified.

Expansion is not the default reward for success. It is a new authorization decision.

Binds through: STD-07 §2.2 ceilings, which bound what a delegation may do without bounding when it may be widened; STD-08 Part A makes expansion its own authorization decision, with its own evidence and its own correction-capacity check; state_history on the grant with expansion transitions; MEC-19 expansion review; Expansion eval, whether scope growth was a decision or a drift.

Law XII. No system may erase the conditions of its own contestability

Institutions question, replace, and withdraw systems using capacities: staff who retain the expertise, processes that still work without the system, records that show what it did, alternatives that remain viable. A consequential system may draw on those capacities. It may not consume them.

Some capacities are preserved precisely because the institution hopes never to need them. The law names them, records them, and treats their erosion as a governance event rather than an efficiency gain.

A consequential system may not consume the institutional capacity required to question, replace, or withdraw it.

Binds through: STD-06 amendment, preserved capacities; preserved_capacities on the dependency record; MEC-15 withdrawal rehearsal and MEC-18 dependency ledger; Contestability Preservation eval.

The variables are not a checklist. They are the coordinates a delegation has at any moment, and the laws are constraints on how those coordinates may move relative to one another.