A digital inspection can record a dimension accurately and still lead a manufacturing team to the wrong decision. If a part is checked against an obsolete drawing revision, the instrument, software and report can all look correct while the conclusion is invalid. Gleb Tsipursky, a behavioral scientist and CEO of Disaster Avoidance Experts, argues in an essay that AI-assisted inspection makes this failure mode more consequential and that manufacturers should attach a version passport to every consequential measurement.
The underlying technical opportunity is persistent identification. Quality Information Framework data can preserve the identity of a measured characteristic across product definitions, inspection plans and results. That supports traceability. The management problem is narrower and harder: each decision must follow the correct version of the requirement. Connected metrology increases the distance and speed at which a result travels, so missing revision context can be carried farther before anyone notices.
Tsipursky's concern is not that AI will calculate badly. It is that an AI tool may receive a result without the version context needed to interpret it. The assistant can summarize a report, point to an apparent deviation or recommend cases for review, but it may compare an observation with the wrong specification. The arithmetic can be correct while the conclusion is invalid. The National Institute of Standards and Technology's 2026 smart-manufacturing roadmap, as cited in the essay, identifies integration of heterogeneous data and trustworthy operation as continuing challenges for industrial AI. A version passport addresses a specific part of that problem by making a measurement's origin and applicable conditions visible to the next decision-maker.
The proposal is deliberately modest. A version passport need not be a new software platform. It can begin as a set of fields that travel with a measurement: the approved drawing or model revision, the characteristic identifier, the inspection-plan revision, the instrument or method, the result timestamp, and the identity of the person or system that approved the disposition. Those fields do not by themselves prove that a part conforms. They establish which requirement was in force when the result was interpreted and who accepted that interpretation.
The fields are not a substitute for a controlled document system, and the essay does not present them as one. Their purpose is to make the measurement self-describing at the point of use. A result that arrives without them asks the next system to infer context that may no longer be available. In a connected workflow, that inference can become invisible: the assistant sees a number, a tolerance and a part identifier, but not the revision history that determines whether the tolerance applies. The passport forces that context into the open without requiring every legacy system to be replaced at once.
Quality teams can test the idea before integrating a new platform. The essay recommends selecting a small set of completed inspection records: one legitimate out-of-tolerance result, one conforming part and one example with mismatched revisions. An AI assistant is asked to summarize the evidence, identify the applicable revision and flag missing provenance. Its responses are then compared with an experienced quality engineer's determination. The test should not stop at how often the assistant labels a part good or bad. More revealing questions are whether it catches a revision mismatch, whether it invents an unstated tolerance and whether it can point to the approved source behind a conclusion.
Those errors are easy to conceal in digital handoffs because each system can appear internally consistent. A result can be formatted cleanly, a report can be complete and a recommendation can sound confident while the reference behind it is stale. The essay therefore argues for clear ownership when records disagree. A quality engineer may determine whether a measurement is valid; design engineering confirms the governing definition; manufacturing determines whether production can continue. The escalation route should be settled before an alert appears, not negotiated during a shipment dispute.
The argument differs from the familiar demand for more comparisons between AI and human inspectors. Even perfect agreement between a human and an automated inspector can reinforce the wrong decision if both were given an obsolete specification. The reference itself needs scrutiny. Automation can make measurement data travel farther and faster, but a reliable chain of custody must ensure that the drawing, feature, method and result still belong together when the decision is made.
That modesty is also a limitation. A set of fields cannot resolve conflicting definitions by itself, and it does not tell a manufacturer which department has authority. It only makes the conflict easier to detect. The ownership rules and the escalation path remain organizational decisions. The essay's test is designed to expose those decisions early, using a small sample rather than a full platform rollout.
For manufacturers, the practical value of the passport is not a new dashboard. It is a discipline of keeping revision, method, timing and approval attached to the measurement as it moves from instrument to report to decision. The proposal does not eliminate the need for engineering judgment, nor does it guarantee that AI will interpret every record correctly. It sets a minimum condition for trustworthy interpretation: the next decision-maker should be able to see which requirement the measurement was meant to satisfy.
Tsipursky concludes that the next generation of connected inspection should make that connection impossible to ignore. The version passport is a small artifact, but it targets a large risk: the possibility that accurate measurement data are used against the wrong version of the truth.