Seminal
The magazine and atlas of EU-funded researchA new publication from the web and AI team at Scott Charnley

Atlas / Italy

European Research Council grant · Running

NOTARI-AI Automating Structured Historical Knowledge from Digitised Archives

Coordinator
Universita Degli Studi Di Torino, Italy
Runs
1 November 2026 to 30 April 2028
Status
Running
EU funding
€150k
Total cost
—
Partners
1 organisations in 1 countries
Type
European Research Council grant
Grant agreement
101335609

What the project is doing

In the project's own words, from its summary on CORDIS.

Archives are living through a paradox: unprecedented investment in digitisation has made historical records accessible and readable–but not analytically usable without significant expert effort and expense. Platforms now convert millions of handwritten documents into searchable text–opening archival records spanning centuries to potential users from genealogists and economists to environmental researchers and policymakers.

Yet the entities and relationships that give these records their full analytical value–the people, roles, properties, and events across archives–remain locked inside the transcription, with no reproducible method for extracting them at scale. NOTARI-AI addresses this bottleneck as a semi-autonomous agentic AI system: rather than processing documents through a fixed pipeline, it deploys mutually reinforcing reasoning processes that interrogate, cross-check, and recursively refine one another–grounded in a verified knowledge base and doing computationally what a skilled expert archivist does manually, reading each document against its people, places, and institutions.

A human-in-the-loop validation cycle repositions expert labour from manual encoding to high-level supervision, with every extracted data point retaining an explicit, auditable link to its source. The Proof of Concept will move NOTARI-AI from demonstrated feasibility to institutional readiness: testing performance and transferability across committed state and private partner archives, formalising the validation workflow, and generating evidence for a credible deployment pathway for archives, research programmes, and heritage institutions that have invested in digitisation but cannot yet use what they hold.

The result will be the first reproducible, transferable architecture for converting digitised archival text into verified relational knowledge–and a demonstration that agentic AI can be deployed responsibly where precision, traceability, and accountability are non-negotiable.

Who is involved

OrganisationCityCountryRole
Universita Degli Studi Di TorinoTorinoItalyCoordinator

Source: CORDIS, European Commission. Data of 5 Nov 2026, reused under the Commission's reuse policy. View on CORDIS · Grant DOI