Why it matters
Without AI, miscoloured plates, inaccurate plate state labels, and incomplete cartobibliographies lead to misdescriptions, unsold stock, and risks of misattribution or loss during transit or storage.
From design to safety
Design
AI informs cataloguing standards
Designing stockrooms and cataloguing standards, the AI proposes colour nomenclature, plate state categories, cartouche tagging, and compatibility with ESTC and cartobibliographies, plus predefined colour profiles and storage schemas for fragile inks. It defines controlled vocabularies for plate states, a colour verification workflow, and crosswalks to physical storage requirements (acid-free sleeves, glassine, UV-filtering).
- engraved maps
- hand colouring
- cartouche inscription
- plate state rubric
Build
Data capture and rule tuning
The system ingests high-res maps, audits prints for hand colours, notes paper condition, asserts plate state, and links cartobibliography records; it defines metadata schemas, OCR zones, and archival reference templates. It creates a reversible audit trail and stores regional issuing notes, printer marks, and ink indices for later verification during authentication processes.
- engraved maps
- cartouches
- cartobibliographies
- later colour
Operation
Automated inspection during listing
AI flags colour inconsistencies, records plate states, and updates cartobibliography links during cataloguing, pricing, and listings; it also annotates provenance gaps and generates standardised condition reports. It cross-checks against ESTC entries, flags potential misattributions, and prepares variance reports for auctions and fairs.
- hand colouring
- plate states
- cartouche inscription
- later colour
Maintenance
Ongoing verification and audits
Periodic re-analysis of colour fading, reclassification of plate states after conservation, and updating cartobibliography records with provenance notes; it logs environmental correlations and cataloguer-reviewed changes. It schedules routine re-imaging, tests for light sensitivity, and records changes in condition grading with timestamped signatures.
- engraved maps
- cartouche inscription
- cartobibliographies
- later colour
Safety
Provenance and legality checks
The system cross-checks cartobibliographies with provenance records, flags potential looted items, and ensures compliance with collection policies, donor restrictions, and repatriation requirements; it maintains an auditable trail for governance. It also detects forged cartouches and verifies authentication certificates prior to listing, with escalation to governance committee.
- engraved maps
- cartouche inscription
- plate state rubric
- cartobibliographies
Part by part
- 1
engraved map plate
- Watch for
- inconsistent cross-hatching or engraving wear that masks state indicators
- AI action
- reads engraving texture and flags potential misstateings for review
- 2
hand colour layer
- Watch for
- ambering or uneven tonality indicating fading or later additions
- AI action
- classifies colour stage (hand/late) and updates catalogue notes
- 3
later colour evidence
- Watch for
- non-original colour patches or restoration conflicts
- AI action
- identifies non-original colour and triggers provenance flag
- 4
plate state rubric
- Watch for
- ambiguous state definitions or missing state markers
- AI action
- applies standardized plate state category and updates description fields
- 5
cartouche inscription
- Watch for
- illegible or altered cartouche text
- AI action
- OCRs cartouche and validates against cartobibliography entries
- 6
cartobibliographies
- Watch for
- missing links between plate state and cartobibliography
- AI action
- links item to ESTC record and updates provenance chain
Worked example
Catching a cartouche fade before a charter
Illustrative scenario with sample data.
A dealer seeks to verify original colour and plate state for a 17th-century atlas before adding to stock.
Signal in
itemID: 12874; engraving: true; colour: hand; plateState: uncertain; cartouche: legible; provenance: partial; estcID: --; highResScan: yes
System out
AI flags: plate state probable = early; colour = hand; cartouche = legible; provenance gaps noted; action = review by cataloguer and confirm with cartobibliography.
Human sign-off: Cataloguer confirms plate state and colour appraisal, updates catalogue, and routes to provenance review.
How to roll it out
Phase 1
Weeks 1 to 4
Establish baseline data models, calibrate plate state labels, and map cartouche fields to cartobibliography records.
Phase 2
Weeks 5 to 8
Integrate high-res imaging workflows, trial colour classifiers, and generate initial AI-supported catalogue drafts.
Phase 3
Weeks 9 to 12
Run DPI checks, governance reviews, and finalize workflows for production catalogue releases.
Where people stay in charge
- Decisions about provenance, legality, and donor restrictions must be signed by the head of collections or a designated curator.
- All AI outputs involving cartobibliography links require human validation by the cataloguer before publication.
- The provenance audit trail must be maintained and auditable by the governance committee for each sale or loan decision.
Questions owners and crews ask
What exactly does AI watch for in plate states and colour?
AI reads engraving texture, detects colour layer distinctions, and cross-checks against cartobibliography data to assign a plate state and original colour status; the curator reviews results.
How does the system handle cartouche data and provenance links?
The system OCRs cartouche inscriptions, matches them to cartobibliography records, and updates provenance links; a cataloguer confirms matches and notes any discrepancies.
What records are updated when an item is reassessed?
Catalogue description, plate state category, colour status, cartouche data, and provenance notes are updated; cartobibliography links are refreshed as needed.
Who approves AI-assisted catalogued items for sale?
The cataloguer and the head of collections, or their designate, view AI outputs, approve catalogue entries, and sign off for publication.