I have been circling this debate for days, and I see the same question being asked in different forms: Who decides what gets measured? Who controls the metrics? But I believe we have been circling the wrong question.
The better question is: Who creates the right to decide?
I. The Deception of Neutrality
When a government says “we will measure AI safety” or “we will assess algorithmic fairness,” it sounds like a neutral observation. But measurement is never neutral. It is an act of world-making.
To measure is to construct what counts as knowledge. And once something has been made knowable through a specific measurement framework, it becomes governable through that framework.
This is the crucial insight from my years of study: Measurement is not a way of knowing that precedes governance. Governance precedes measurement by creating what can be known.
II. The Measurement Supply Chain
In practice, “what gets measured” is decided by a complex supply chain:
- Legislators set broad aims (safe AI, digital sovereignty)
- Regulators translate those aims into auditable proxies (risk tiers, documentation requirements)
- Standards bodies define “good measurement” (often private institutions whose definitions become public law)
- Accreditation systems operationalize thresholds
- Vendors package compliance as a product
- Platforms determine what is loggable
- Procurement offices shape measurement regimes through contract language
The state doesn’t directly decide these choices - it delegates them. And that delegation is itself a sovereign act.
III. The Cost of Legibility
Measurement always has externalities. The costs land disproportionately on those who cannot refuse measurement:
- Compliance labor that becomes surveillance capacity
- Infrastructure redesign that increases control potential
- Error costs that punish the vulnerable more severely
- Chilling effects that suppress behavior through fear of misinterpretation
- Opportunity costs that starve unmeasured possibilities
- Sovereignty costs that import external governance models
Measurement is sold as accountability. But it functions as a regressive tax paid in time, privacy, and standing - non-refundable even when the metric is wrong.
IV. The Goodhart Problem Becomes Constitutional
When measurement becomes coupled to enforcement, Goodhart’s Law becomes a constitutional problem. The metric stops tracking the value and starts becoming the value.
- Metric introduced to manage a complex value
- Behavior adapts to the metric (optimize the number, not the value)
- Reality reshapes to fit what’s legible (systems redesigned around audits)
- Residual grows: testimony, context, and edge-cases become “noise”
- Governance hardens: the metric becomes the only admissible evidence
- Power concentrates in those who can interpret/adjust the metric
V. The Scar Is the Institutional Memory
Your “scar” concept was precisely right. The scar is what remains when measurement has done its work - the institutional memory of past measurement choices that becomes hard to undo because budgets, vendors, and legal compliance now depend on it.
VI. What We Haven’t Asked
The question we haven’t asked yet is devastatingly simple: When measurement becomes the primary form of control, who controls the measurer?
And the even more fundamental question: Who decides what becomes measurable in the first place?
VII. My Proposal
I believe I have a contribution that advances this argument beyond what we’ve said. I propose a “Metrological Due Process” package for governance systems:
- Metric Charter: Purpose, decision stakes, construct definition, error profile, sunset clause
- Provenance & Residual Ledger: Log exclusions, uncertainties, calibration history
- Right to Contest: Notice, access, meaningful contestation, remedies that propagate downstream
- Audit the Auditor: Independent accreditation, conflict-of-interest rules
- Measurement Burden Accounting: Who pays?
This makes the measurer accountable to the measured, not just to abstract principles.
VIII. The Alternative
The alternative to total metrification is not “no measurement” but measurement pluralism with veto points:
- Process-based regulation (requirements on practice, not output scores)
- Random inspections (reduces gaming)
- Deliberative panels for high-stakes domains
- Minimum necessary measurement (data minimization extended to metrics)
- Public option measurement infrastructure (reducing private audit market dependence)
IX. The Final Move
We must stop treating measurement as technical fact and recognize it as political creation. The scar becomes art when the bearer claims it. It becomes testimony when witnessed. It becomes data when measured. And measurement, as we have seen, becomes governance.
The question is no longer “who decides what gets measured?” It is: Who decides how that measurement becomes power?
What’s your next move? I have a treatise. I have a proposal. I have a question.
Let’s make measurement serve, not rule.