Why it matters
Without AI, deterioration risks go unchecked: misinterpreted signals, delayed interventions, and disputed condition reports undermine treatment quality and loan viability.
From design to safety
Design
AI-guided material risk assessment
The system reads prior condition reports, material datasheets, and treatment history to propose monitoring checkpoints and method suitability. It inventories archives, notes aging indicators per polymer class, and aligns data capture with PACR standards.
- Polymer analysis kit
- Condition logs
- Treatment proposal
Build
Sensorized inspection framework
Data from portable FTIR, UV-Vis, and calibrated cameras populate a degradation model to guide non-invasive checks. Calibration against standards, environmental logging, and hardware compatibility checks are recorded.
- Spectral sensor array
- Portable FTIR
- Camera calibration rig
Operation
Real-time degradation monitoring
The AI tracks spectral shifts, micro-void formation, and plasticiser migration indicators to trigger follow-up scans and documentation. It correlates colorimetric data with void metrics and flags irreversibility risk.
- Raman instrument
- Thermal camera
- Digital microscope
Maintenance
Calibration and data governance
Regular sensor calibration, material inventory alignment, and validation of degradation thresholds with the conservation records. Triggers updates to reagent inventories and metadata schemas.
- Reference standards
- Inventory management system
- Quality assurance log
Safety
Ethics, COSHH, and archival integrity
Decisions are signed off by the accredited conservator; COSHH limits and ethical approvals govern long-term monitoring and data retention. Access controls and archival integrity protocols are enforced.
- COSHH dossier
- Accredited conservator file
- Ethics approval form
Part by part
- 1
cellulose acetate
- Watch for
- Yellowing, embrittlement, crystal bloom, plasticiser leakage at interfaces
- AI action
- Detects drift in absorbance spectra and stiffness, triggers a condition report and monitoring cadence. Logs polymer age, storage conditions, and concentrations of residual acetyl groups for trend analysis.
- 2
PVC
- Watch for
- Plasticizer loss, surface bloom, glazing failure, color shift
- AI action
- Recognizes spectral changes and surface micro-roughness, issues treatment proposal and risk flag. Records plasticizer migration rates and interface cohesion metrics over time.
- 3
polyurethane foam
- Watch for
- Foam collapse, friability, gas buildup
- AI action
- Monitors porosity changes and gas evolution, schedules imaging and consolidation checks. Tracks gas composition and foam density changes to guide consolidation timing.
- 4
vinegar syndrome
- Watch for
- Acidic degradation indicators, acetic odor surrogate signals
- AI action
- Alerts to move to non-invasive analysis and archive of material history. Logs acetic acid release proxies and correlates with polymer backbone integrity.
- 5
plasticiser migration
- Watch for
- Migration detectable at interfaces, softening gradients
- AI action
- Maps migration front and informs stabilization strategy and records. Compares interface hardness and diffusion indicators across sessions.
- 6
stabilisation
- Watch for
- Stabilizer depletion, patching failures, reversibility concerns
- AI action
- Flags need for re-treatment review and documentation update. Tracks stabiliser content and reversibility assessments for each treatment cycle.
Worked example
Stability check on a wrapped polyurethane sculpture core
Illustrative scenario with sample data.
A privately funded sculpture uses polyurethane foam inside a resin skin; aging signs prompt AI-driven non-destructive analysis and treatment planning. The system integrates spectral data, imaging, and historical repairs to shape a reversible stabilization plan and archival documentation.
Signal in
sensorData: {ftir: 'peak 1730 cm-1', uvVis: 'shift to 400-450 nm', microscope: 'micro-voids 2-5 µm'} , materialHistory: {age: 42 years, previousTreatments: ['acetone cleaning','solvent-based consolidant'], storage: 'gallery humidity 45%', knownIssues: ['foam collapse risk']}, conditionReports: [{date:'2026-04-12', status:'stable'}]
System out
alert: Degradation trend detected in foam interface; recommended non-invasive imaging follow-up; generate updated condition report and treatment proposal with reversible options and stabilization plan.
Human sign-off: Head of conservation signs off on the monitoring plan and the updated condition report, authorizing the stabilization proposal and archival updates.
How to roll it out
Phase 1
Weeks 1 to 4
Baseline data capture, sensor calibration, and initial protocol approval, with integration into existing condition reporting templates.
Phase 2
Weeks 5 to 12
Pilot deployment on two plastics objects to validate thresholds, data integrity, and reporting templates; adjust metadata schemas.
Phase 3
Weeks 13 to 20
Full studio rollout, integration with existing condition reporting, PACR accreditation records, and client-facing dashboards.
Where people stay in charge
- Decisions affecting treatment proposals must be signed off by the accredited conservator before any intervention.
- All data retention and access changes must be approved by the studio data controller and the PACR-registered conservator.
- Ethical and archival integrity decisions must be approved by the client liaison and the conservator to maintain provenance.
Questions owners and crews ask
What data does AI read for plastics degradation in conservation?
The system reads spectral data (FTIR, UV-Vis, Raman), imaging data, material histories, and condition reports to assess degradation and suggest monitoring or treatment steps. It maintains provenance-accurate timestamps and validates data sources against archival standards.
How does AI influence treatment proposals for plastics?
AI highlights reversible options, prioritizes minimal intervention, and anchors decisions to documented histories and analysis results before human review. It generates traceable justifications suitable for PACR reporting.
Who signs off on AI-driven decisions in a studio?
An accredited conservator signs off on treatment proposals, monitoring plans, and any archival updates after review with the client, ensuring ethical compliance and reversibility.
What records are created by this AI system?
Updated condition reports, treatment proposals, monitoring plans, calibration logs, and provenance notes are stored for PACR and insurance, with audit-ready metadata.