AI Coding Assistant for COBOL Development

Employment & Work

What it collects

Operational data

Anonymized data

Internal backend mainframe COBOL source code from CRA systems. This is operational/technical code, not personal information about taxpayers or individuals.

Government of Canada AI Register — COBOL Coding Assistant

Run by: Canada Revenue Agency (CRA)
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

Employment & Work

This AI tool helps Canada Revenue Agency software developers write, document, and test COBOL code — a legacy programming language still central to federal tax systems. It generates code snippets, detects syntax errors, and produces documentation automatically. Only government employees use this tool directly; it does not process personal information about Canadians.

What it collects and what happens to it

Data taken in

Operational data

Anonymized data

Internal backend mainframe COBOL source code from CRA systems. This is operational/technical code, not personal information about taxpayers or individuals.

Government of Canada AI Register — COBOL Coding Assistant

Processing

Language Models

Built on open-source generative AI / large language model technology, fine-tuned or prompted to understand and generate COBOL programming language syntax and patterns.

Government of Canada AI Register — COBOL Coding Assistant

Anomaly Detection

Detects syntax errors and coding mistakes in submitted COBOL code, flagging deviations from correct patterns and consistent coding practices.

Government of Canada AI Register — COBOL Coding Assistant

What it does

Creating (Generative AI)

Human decides

Generates COBOL code snippets, documentation, and test cases in response to developer prompts. Developers review and decide whether to accept, modify, or discard the generated output.

Government of Canada AI Register — COBOL Coding Assistant

Deciding (Analytical AI)

Human decides

Analyzes existing COBOL source code for syntax errors and coding issues, returning a classification or diagnostic that a human developer then reviews and acts upon.

Government of Canada AI Register — COBOL Coding Assistant

Outputs

Generated content

Anonymized data

Generated COBOL code snippets, technical documentation, and test cases. These are software artifacts, not personal data. Developers review all generated output before use.

Government of Canada AI Register — COBOL Coding Assistant

A recommendation or prediction

Anonymized data

Syntax help and coding error diagnostics presented as advisory suggestions to the developer. These recommendations are non-binding — the human developer decides whether and how to apply them.

Government of Canada AI Register — COBOL Coding Assistant

Run by

Canada Revenue Agency (CRA)

The Canada Revenue Agency (CRA) is the federal institution deploying this COBOL coding assistant to support its software development teams who maintain legacy mainframe tax systems.

Government of Canada AI Register — COBOL Coding Assistant

Built by

Not stated by the Helpful Places.

Kept for

Not stated by the Helpful Places.

Shared with

Not available to me

This tool is an internal developer productivity assistant restricted to GC employees. Members of the public have no access to its outputs or the underlying CRA mainframe code it processes.

Available to the accountable organization

Generated code, documentation, and diagnostic outputs are available to authorized CRA developers and their supervisors within the agency's development environment.

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 Be Informed of AI UseGC employees using this tool are informed that AI is in use — the register confirms AI use is disclosed to users. Members of the public are not direct users and are not subject to AI-mediated decisions through this system.
Right to Algorithmic TransparencyThis AI system is listed on the Government of Canada's public AI register, providing transparency about its purpose, data sources, and capabilities. The register entry is publicly accessible at open.canada.ca.

Risks and safeguards

Reputational harmthe AI may generate incorrect COBOL code or faulty documentation that, if accepted without review, could introduce bugs into tax systems relied upon by millions of Canadians.Safeguard: the system operates in a human-decides mode — all generated code must be reviewed and approved by a developer before integration. The tool is in development status, allowing for staged rollout and quality validation before broad deployment.
Societal & cultural harmover-reliance on AI-generated COBOL code could further erode the pool of human COBOL expertise within government, creating long-term dependency on AI tools for critical tax infrastructure.Safeguard: the register frames the tool as augmenting — not replacing — developer skills, targeting productivity and quality improvements while developers retain decision authority over all outputs.