AI Risk Assessment Tool for Automated Decision-Making

Risk Assessment & Triage · Planning & Decision-making

What it collects

Operational data

Anonymized data

Questionnaire responses describing the proposed automated decision system under assessment — including information about the system's design, the algorithm being used, the type of decisions it makes, and its potential impacts. Input data is about the system being assessed, not about specific individuals.

Algorithmic Impact Assessment — Government of Canada Open Data

Run by: Treasury Board of Canada Secretariat (TBS)
Where: No fixed location
Kept: Not stated by the Helpful Places.
Shared with: Accountable organization, Download

What it is for

Risk Assessment & Triage

Planning & Decision-making

The Algorithmic Impact Assessment (AIA) is a mandatory questionnaire tool used by the Government of Canada to evaluate the potential risks of automated decision systems before they are deployed. It scores systems across dimensions like data quality, algorithm design, and decision impact, assigning an impact level that determines required safeguards. Any federal department deploying an automated decision system must complete this assessment. The tool is publicly available under an open license to support transparency and broader adoption of responsible AI practices.

What it collects and what happens to it

Data taken in

Operational data

Anonymized data

Questionnaire responses describing the proposed automated decision system under assessment — including information about the system's design, the algorithm being used, the type of decisions it makes, and its potential impacts. Input data is about the system being assessed, not about specific individuals.

Algorithmic Impact Assessment — Government of Canada Open Data

Processing

Classification & Prediction

The AIA uses a rule-based scoring algorithm to classify an automated decision system into one of four impact levels (I through IV). Assessment scores are computed from weighted responses across risk dimensions including system design, algorithm type, decision type, and potential impact on affected parties.

Algorithmic Impact Assessment — Government of Canada Open Data

What it does

Deciding (Analytical AI)

Human decides

The AIA questionnaire scores and classifies a proposed automated decision system into one of four impact levels based on responses to risk and mitigation questions. The resulting score is advisory — it informs required governance actions that departmental officials then decide how to implement.

Algorithmic Impact Assessment — Government of Canada Open Data

Outputs

A recommendation or prediction

Anonymized data

The AIA produces an impact level score (I–IV) and corresponding recommended governance requirements for the assessed automated decision system. This output is advisory — it guides departmental compliance with the Directive on Automated Decision-Making but does not itself constitute a binding government decision about an individual.

Algorithmic Impact Assessment — Government of Canada Open Data

Run by

Treasury Board of Canada Secretariat (TBS)

The Treasury Board of Canada Secretariat developed and maintains the Algorithmic Impact Assessment tool. It is the accountable federal body responsible for the Directive on Automated Decision-Making under which the AIA is mandated.

Algorithmic Impact Assessment — Government of Canada Open Data

Built by

Not stated by the Helpful Places.

Kept for

Not stated by the Helpful Places.

Shared with

Available to the accountable organization

Assessment results and questionnaire responses are available to the federal department completing the AIA and to the Treasury Board of Canada Secretariat for compliance oversight purposes.

Available to download

The AIA tool, its questionnaire, and its source code are publicly available for download and reuse under an open license, consistent with the Government of Canada's commitment to open government.

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 AIA tool itself is available to the public under an open license. Its source code repository is publicly accessible, and the scoring methodology is transparent. Members of the public can access the AIA questionnaire and understand the criteria used to evaluate automated decision systems deployed by the Government of Canada.
Right to Be Informed of AI UseThe Directive on Automated Decision-Making, which the AIA supports, requires federal departments to notify individuals when an automated decision system is being used to make or assist in decisions that affect them. Completing the AIA is a prerequisite to deployment, making this notification requirement enforceable through the compliance process.
Right to a Human ReviewThe AIA scoring methodology explicitly evaluates whether proposed automated decision systems include meaningful human review pathways. Systems assessed at higher impact levels (III and IV) under the Directive are required to have human oversight mechanisms, including the ability for affected individuals to have decisions reviewed by a human official.

Risks and safeguards

Civil liberties harmIf the AIA impact classification is miscalibrated or incomplete, automated decision systems with serious civil liberties implications (e.g. those affecting due process, access to services, or freedom from discriminatory treatment) may be deployed with insufficient safeguards.Safeguard: The AIA was developed in consultation with internal and external stakeholders and is maintained as open-source software to allow public scrutiny. The Directive on Automated Decision-Making mandates escalating safeguards at higher impact levels, including human review requirements and notice to affected individuals.
Loss of autonomyDownstream automated decision systems that receive low impact scores from the AIA may be deployed with minimal human oversight, reducing individuals' ability to contest or understand decisions that affect them.Safeguard: The AIA's scoring methodology includes specific questions about the availability of recourse mechanisms, opt-out options, and human review pathways; systems lacking these protections receive higher risk scores that trigger more stringent governance requirements.