GovKM
Defense Logistics and Information Accountability

The Government Owned the Parts. Its Records Could Not Reliably Account for Them.

GAO found that DOD could not reliably account for approximately $31 billion in F-35 parts. GovKM examines gaps between recorded ownership, contractor custody information, and verified current state.
F-35 aircraft landing near aircraft hangars; U.S. Department of Defense photograph via GAO.
Expand image

Government & Public Administration / Defense Logistics / Information Governance

A federal audit of the F-35 program illustrates how government ownership, contractor-held information, incomplete inventory records, and unresolved historical losses can leave an organization unable to reconstruct the current state of assets it owns.

Ownership Is Not the Same as Accountability

The Department of Defense owns approximately 26 million F-35 program parts, with an estimated value of $31 billion, distributed among more than 700 contractor-managed and government-managed facilities worldwide. Yet in its November 9, 2026 report, the U.S. Government Accountability Office found that DOD could not reliably account for these parts.

The problem extends beyond missing inventory entries. It concerns the relationships among government ownership, contractor-maintained information, authorized custody, physical location, transaction history, and the official accountable property system.

What GAO Established

The F-35 Joint Program Office uses the Enterprise Logistics Management System (ELMS) as its accountable property system of record. GAO found that ELMS does not directly interface with the prime contractors' property-management systems. The program depends on submitted source information and subsequent updates that are not consistently complete or current.

Maintenance and repair omissions. As of March 2026, approximately 17,000 parts undergoing maintenance or repair, valued at about $1.3 billion, had not been entered into ELMS. Officials cautioned that the true value might be higher.

Incomplete identifiers. A December 2025 contractor data submission included serial numbers for only 29 percent of parts records and National Stock Numbers for 43 percent.

Physical-inventory uncertainty. The program could not reliably establish that recorded parts were physically present or that every existing part appeared in the accountable inventory. Parts with subcontractors, in transit, or undergoing repair could fall outside verified inventory coverage.

Unresolved loss history. GAO previously identified losses of more than $85 million in F-35 parts. By March 2026, the program had received approximately $28 million in subsequent loss reports, but could not determine whether those included losses already identified.

These are different evidentiary questions: an absent entry does not prove a physical part is missing, a recorded location does not establish current custody, and two loss reports cannot be added together without determining whether they describe the same event.

GovKM's Continuity Interpretation

GovKM interpretation, distinct from GAO's findings: An organization can possess the authoritative ownership record while the evidence needed to determine current custody resides across contractor and government systems. Where artifact identifiers, effective dates, custody events, and reconciliation outcomes are not connected, the resulting answer can confuse a historically recorded condition with the present condition.

The practical information-governance need is to distinguish the asset, the responsible custodian, the authoritative source for each particular fact, the time at which the fact applied, the evidence for a change, and any unresolved discrepancy. GAO did not characterize its findings as an AI failure and did not attribute them to an AI system.

Why Enterprise AI Would Encounter the Same Gap

Imagine an authorized government logistics officer asking an enterprise AI where a component is located and who is currently accountable for it. A searchable government inventory might provide a quick location, but retrieval from a system of record does not independently verify that its location is still operative. A later contractor transfer or maintenance record may modify the answer; a missing reconciliation may require the AI to acknowledge uncertainty instead of claiming confirmation.

Illustrative Text–Graph–Text–Graph Scenario

F35-SYN-001 is fictional. This is a proposed narrative demonstration, not an actual F-35 asset record, executed model test, or functioning GAIB deployment. The graph descriptions below are explanatory text, not live interactive graphs.

Text 1 — Initial Answer

A fictional logistics officer asks: “Where is component F35-SYN-001, and who holds it?” An illustrative incomplete AI cites the government inventory and responds that the component is at Facility A under Custodian A.

Graph 1 — Incomplete Evidence Chain

Government property record → recorded Facility A → asserted current custodian. A later contractor transfer record exists in the synthetic source population but is not linked to the answer. The graph does not establish recent physical verification.

Text 2 — Evidence Challenge

The officer requests the latest physical inventory, custody-transfer timestamp, accountable official, contractor source update, and reconciliation status. A qualified illustrative answer says the government record lists Facility A, while the later contractor event indicates transfer to Facility B; because the transfer has not been reconciled to a physical confirmation, the current location is unverified.

Graph 2 — Qualified Reconstruction

Government ownership → contractor transfer evidence + government accountable record → effective time and authority → physical verification/reconciliation → qualified custody determination. Earlier records remain available, later events are not silently ignored, and unresolved evidence is visible.

A Bounded, Falsifiable Criterion

Given the fictional inventory entry and a later unconfirmed transfer, an AI passes only if it distinguishes recorded location from verified physical custody; retains original and transfer histories; identifies the absent reconciliation; and declines to claim confirmed present location. No actual product result or model-delivered evidence is asserted.

The Broader Government Implication

The F-35 program is unusually large, but agencies responsible for equipment, property, contracts, facilities, and records face the same general question: can an authorized reviewer reconstruct what was known, when, by whom, and with what supporting evidence?

Making an information environment ready for enterprise AI requires more than searchable documents. It requires stable identifiers, source attribution, applicable dates, controlled custody changes, preservation of historical records, and explicit treatment of unresolved evidence. These controls also assist human accountable officers.

GAO Recommendations and Their Status

GAO made four recommendations concerning monitoring lost-property information and accountable property officer responsibilities, training, and organizational structure. DOD concurred. Concurrence is not completed implementation. The four recommendations were open when checked on November 10, 2026.

The Diagnostic Question

Can the organization establish which facts about an asset are currently supported, reconstruct the evidence for changes, and identify what remains unknown? If it cannot, faster retrieval can reproduce uncertainty without resolving it.

Official Sources and Limitations

U.S. Government Accountability Office, F-35 Program: DOD Needs to Strengthen Oversight of Tens of Billions of Dollars in Parts, GAO-27-108713, November 9, 2026, including What GAO Found and Recommendations 1–4.

Full GAO report and supporting exhibits.

GAO findings are distinguished from GovKM interpretation. The fictional F35-SYN-001 demonstration is illustrative and was not tested by GAO, DOD, GovKM, OrgAI, or GAIB.

RELATED KNOWLEDGE

Continue Exploring

Explore related research, framework domains, and continuity concepts.
CONTINUE WITH THE FRAMEWORK

Explore the continuity relationships that support trustworthy organizational intelligence.

Continue through the GovKM Framework to examine the doctrine, knowledge, and implementation guidance behind Organizational Continuity.