This is NOT an official site of the Government of Canada. Click here for the official AI registry.

AI-Assisted Extraction of Immigration Records from Historical Documents

Accountable

Library and Archives Canada (LAC)

Functional Modes

Sensing (Perceptive AI)

Human decides

Purpose

Research & Development

Input Dataset

Sensitive personal information Identifiable data

Algorithm or Model

Computer Vision

Output Dataset

Sensitive personal information Identifiable data

Details

This system used Amazon Web Services Textract to automatically read and index 11,231 pages of historical Canada Gazette immigration records, making them searchable for genealogy research. It extracted structured text and table data from scanned documents and indexed the results using ElasticSearch. The system involved personal information drawn from historical immigration records and was developed by Library and Archives Canada using a vendor-supplied AI service. This system has been retired and is no longer in active operation.