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

Helpful Places

Government of Canada – DTPR AI Register

Back to your search

AI-Assisted OCR Error Detection and Correction for Fisheries Data

Inform

What it collects

Operational data
Anonymized data
  • Area Stream Inspection Logs from Pacific Regions — handwritten or non-machine-readable fisheries inspection records converted to text via OCR. The register confirms no personal information is involved.
Run by
Fisheries and Oceans Canada (DFO)
Where
No fixed location
Kept
Not stated by the Helpful Places.
Shared with
Accountable organization
Your copy
You cannot see the data it holds about you. What you can do

What it is for

Inform

This tool uses large language models to automatically detect and correct transcription errors introduced when handwritten fisheries inspection documents are converted to digital text via optical character recognition (OCR). It is used internally by Government of Canada employees and does not process personal information. The system is still in development and currently the use of AI is not disclosed to users at point of interaction.

What it collects and what happens to it

Data taken in

Operational data
Anonymized data
  • Area Stream Inspection Logs from Pacific Regions — handwritten or non-machine-readable fisheries inspection records converted to text via OCR. The register confirms no personal information is involved.

Processing

Language Models
  • Large language models (LLMs) from open-source origins are used to apply intelligent validation rules, detect OCR-induced transcription errors, propose corrections, and normalize extracted data into standardized formats.

What it does

Deciding (Analytical AI)
Human decides
  • The system classifies and scores OCR-extracted text fields against validation rules, proposing corrections. Human transcribers review and act on the proposed corrections before data enters the database.
Understanding (Semantic AI)
Human decides
  • Large language models are used to understand the meaning and context of OCR-extracted text in order to identify likely transcription errors and normalize data into standardized formats.

Outputs

Operational data
Anonymized data
  • Validated and normalized fisheries inspection data records, with proposed corrections to OCR errors, ready for entry into standardized databases. No personal information is contained in the outputs.
A recommendation or prediction
Anonymized data
  • The system proposes specific text corrections to human transcribers who make final decisions; it does not autonomously write corrected records without human review.

Run by

Fisheries and Oceans Canada (DFO)
  • Fisheries and Oceans Canada is the federal department deploying this AI data validation tool for internal use by GC employees processing Pacific Region fisheries inspection logs.

Government of Canada AI Register — Data Validation Tool

Built by

Not stated by the Helpful Places.

Kept for

Not stated by the Helpful Places.

Shared with

Not available to me
  • The system processes internal government fisheries operational data. The general public does not have access to the data produced by this system.
Available to the accountable organization
  • Validated data outputs are available to Fisheries and Oceans Canada employees for database entry and operational use.

Stored

Not stated by the Helpful Places.

How to read the colours

Can it identify you?

Anonymized data
Data about people with the link to who is broken. Stripped of identifiers, blurred, aggregated, or noised so this system can’t reasonably tie a record back to an individual.
Pseudonymous data
Each person’s data is tied to a token (hash, ID, template) that lets this system recognise the same person across events, but the token itself doesn’t reveal a name. Reidentification is possible with extra information.
Identifiable data
The data either contains a direct identifier (name, address, account name, recognisable face or voice, plate number) or carries a token this system uses to look up legal identity during processing.

Who completes the loop?

Human decides
This mode suggests; a person decides what to do next. The AI is always advisory — a human is in the loop on every decision. Example: a triage tool ranks cases for a clinician who chooses which to see first.
Human executes
This mode decides; a person carries out the result. Example: an optimizer plans the day’s trash-collection routes, and drivers run them.
Autonomous
This mode decides and acts on its own. No person reviews each decision or carries out the resulting action.

Definitions from the DTPR standard. Amber is about your data, violet about who decides. The fuller the shape and the deeper the colour, the more identifying the data or the less a person is involved.

What you can do

Ask about this system

Questions go to the Helpful Places, not the vendor.

Your rights

  • Right to Algorithmic TransparencyThe register records that AI use is not currently disclosed to users at the point of interaction. As an internal government tool, members of the public are not direct users; however, GC employees using the system may not be informed that AI is in operation. This right is not currently operationalized for this system.

Risks and safeguards

  • Reputational harmLLM-proposed corrections to OCR errors could introduce new errors or normalize data incorrectly, compromising the integrity of fisheries inspection records used for regulatory and scientific purposes.Safeguard: The system proposes corrections for human review rather than applying them autonomously; human transcribers retain final decision authority before data enters the database.