Selection Gating and the Certified Basin
Airworthiness certification paired with the near-universal requirement of insurance produces a genuine selection basin: an aircraft that fails certification cannot be deployed, because no insurer will cover it and no airport will schedule it.
What decision changes?
Assemble procurement, insurance, and licensing levers jointly rather than relying on any one of them — alignment is selected, not merely designed, only when the uncertified path is made more expensive than the certified one across all three.
Evidence and authority are not enough if uncertified actors can still operate profitably. What makes aviation safety actually hold is not the certification stamp itself but the lattice of consequences around it: an aircraft that fails airworthiness certification essentially cannot be deployed, because no insurer will cover it, no airport will schedule it, and no airline can legally fly it commercially. Each barrier alone might be evaded; together they make the uncertified path more expensive than the certified one at every turn. A manufacturer does not comply because it is virtuous — it complies because non-compliance has no market.
The same interlocking structure appears wherever certification genuinely bites. An unapproved drug cannot be lawfully prescribed, stocked by pharmacies, or reimbursed by insurers. In nuclear power, the industry’s own peer-inspection body (INPO, founded by the utilities themselves after Three Mile Island) issues ratings with real consequences for a plant’s insurance costs and licensing standing — an industry disciplining itself because a single member’s accident threatens everyone’s license to operate.
One artifact from this world deserves particular notice: the airworthiness directive. It is terse, addressed across organizational roles, and directly decision-changing — a specific aircraft component must be fixed by a specific date or the aircraft is grounded. Compare a typical safety report, which no procurement officer, insurer, or engineer is obligated to act on. The lesson is that safety knowledge only reshapes behavior when it is packaged in a form some gatekeeper is bound to act on.
For AI, the open question is which gatekeepers can be assembled into an equivalent lattice — compute providers’ terms of service, cloud-deployment conditions, government procurement requirements, insurance underwriting — and whether it can be assembled before uncertified deployment gains irreversible market share. History’s warning is precise on this point: certification without the surrounding consequence structure is paperwork, and the consequence structure takes longer to build than the certificate. Safety practice, in the end, is selected, not merely designed.
One of eleven historical case studies in Institutional Genesis, Memory, and Decay — see the overview for the full life-cycle map, or read the complete appendix.
What would count as evidence?
The airworthiness directive is the paradigm conductive artifact: terse, cross-role, and directly decision-changing, unlike a safety report no procurement officer is obligated to act on; the Institute of Nuclear Power Operations (founded after Three Mile Island) shows the same pattern via peer-inspection standards tied to insurance and licensing standing.