Proposed interviewees: A marque historian, John Draneas and the relevant archive representative.
Buyer takeaway: Use automated document review to organise questions; require accountable human verification of purchase-critical conclusions.
Why it matters: The Gooding Gullwing addendum discussed earlier says a registry entry placing the car in Italy in 1994 is believed incorrect. That is a concrete reason to investigate provenance discrepancies rather than assume a registry and a catalogue are independent, mutually confirming evidence. (goodingco.com)
NIST identifies confidently presented false information and fabricated supporting citations as generative-AI risks. This is general technology evidence, not an evaluation of any particular automotive authentication product. (nvlpubs.nist.gov)
Put this question to the expert: “When a title, registry, restoration file and auction catalogue disagree, how do you identify the underlying source and decide which claim remains supportable?”
A good answer should cover:
- An event-by-event chronology, including explicit gaps.
- Original documents versus copies, transcriptions and later summaries.
- Whether apparently separate accounts trace back to the same owner’s statement.
- Permissioned checks with archives, workshops and prior custodians.
- Physical evidence that can corroborate or contradict the paper history.
AI boundary proposed by the desk: Permit extraction, comparison and contradiction flags; require page references and human checking. Do not issue an authenticity verdict from a scan alone or treat an AI-generated citation as evidence.
Buyer deliverable: A claim-to-source register labelled documented, attributed, inconsistent or unknown, with a named person responsible for resolving each material discrepancy.
Expert quote: [Awaiting interview and approval.]
Sources cited on this page
- goodingco.comgoodingco.com
- nvlpubs.nist.govnvlpubs.nist.gov
