Difference between revisions of "Treasury Board Directive on Automated Decision Making"

From wiki
Jump to navigation Jump to search
m
 
(14 intermediate revisions by 4 users not shown)
Line 1: Line 1:
The Government of Canada is increasingly looking to utilise technology and automated systems to make, or assist in making, administrative decisions to improve service delivery. It is committed to doing so in a manner that is compatible with core administrative law principles such as transparency, accountability, legality and procedural fairness. The proposed ''Treasury Board Directive on Automated Decision-Making'' is being developed to provide federal departments and agencies with guidance on how to responsibly deploy automated decision systems.
+
La version française est disponible [[Directive sur la prise de décisions automatisée|ici]] 
  
Systems are classified into four levels of impact. To ensure that governance of these systems is sensible, requirements of the Directive scale to the level of impact.
+
== Background ==
 +
The [https://www.tbs-sct.canada.ca/pol/doc-eng.aspx?id=32592 Directive on Automated Decision-Making] (directive) sets requirements for federal institutions to ensure that the use of AI or other automated systems in making or supporting administrative decision-making is compatible with the core principles of administrative law such as transparency, accountability, legality and procedural fairness. The Directive is supported by the [https://www.canada.ca/en/government/system/digital-government/digital-government-innovations/responsible-use-ai/algorithmic-impact-assessment.html Algorithmic Impact Assessment Tool,] an online questionnaire which helps departments understand and manage the risks associated with automated decision systems.  
  
It will be supported by an Algorithmic Impact Assessment tool, an interactive questionnaire designed to help departments identify their impact level.
+
The directive came into effect in April 2019 and applies to systems developed or procured after April 2020.  
  
[[:File:Treasury Board Directive on Automated Decision Making.pdf|You may find the latest version of the Directive here]].
+
== 4th Review ==
 +
Engagement on the policy recommendations and proposed modifications to the directive as part of the 4th review is underway. To participate, visit our [[Consultations on the 4th Review|Consultations on the 4th review page]] and respond to the survey. '''Consultations are open until January 8, 2025.'''
 +
 
 +
To determine the scope of the 4th review, initial steps involved completing literary reviews, jurisdictional and legal scans for a number of topics in the context of automation and artificial intelligence.  
 +
 
 +
Next steps include:
 +
 
 +
* Drafting recommendations – complete
 +
 
 +
* Targeted consultation and feedback analysis – underway
 +
 
 +
* Governance and approvals – winter 2025
 +
 
 +
* Publication – late spring 2025
 +
 
 +
== Guidance ==
 +
Treasury Board of Canada Secretariat (TBS) has developed guidance to support compliance with the directive:
 +
 
 +
* [https://www.canada.ca/en/government/system/digital-government/digital-government-innovations/responsible-use-ai/guide-scope-directive-automated-decision-making.html Scope] – to support departments in understanding the systems and uses that fall under the purview of the directive
 +
 
 +
* [[:File:Guide on peer review.pdf|Peer review]] – to support departments commissioning a review and individuals undertaking the review
 +
* Section 4.5 of the [https://www.canada.ca/en/government/system/digital-government/guideline-service-digital.html#ToC4_5 Guideline on Service and Digital] provides additional guidance on the responsible and ethical use of automated decision systems
 +
* [[:File:AIA Guide (2024) EN.docx|Algorithmic Impact Assessment Guide]] – to support departments in completing the Algorithmic Impact Assessment and improve the consistency and quality of questionnaire responses
 +
Upcoming:
 +
 
 +
* Reporting – to support departments on the requirement of publishing information on the effectiveness and efficiency of a system in meeting program objectives
 +
 
 +
== Resources ==
 +
 
 +
* [https://www.tbs-sct.canada.ca/pol/doc-eng.aspx?id=32603 Policy on Service and Digital]
 +
* [https://www.canada.ca/en/government/system/digital-government/digital-government-innovations/responsible-use-ai.html Responsible use of artificial intelligence (AI)]
 +
* [[Third Review of the Directive on Automated Decision-Making]]
 +
* [[:File:Treasury Board Directive on Automated Decision Making.pdf|Draft 1.0 of the Directive on Automated Decision-Making]] (2018)
 +
 
 +
== Enquiries ==
 +
Please contact the TBS Responsible Data and AI team (ai-ia@tbs-sct.gc.ca) for any questions.
 +
 
 +
__INDEX__
 +
 
 +
[[fr:Directive sur la prise de décisions automatisée]]

Latest revision as of 13:32, 20 November 2024

La version française est disponible ici

Background

The Directive on Automated Decision-Making (directive) sets requirements for federal institutions to ensure that the use of AI or other automated systems in making or supporting administrative decision-making is compatible with the core principles of administrative law such as transparency, accountability, legality and procedural fairness. The Directive is supported by the Algorithmic Impact Assessment Tool, an online questionnaire which helps departments understand and manage the risks associated with automated decision systems.  

The directive came into effect in April 2019 and applies to systems developed or procured after April 2020.

4th Review

Engagement on the policy recommendations and proposed modifications to the directive as part of the 4th review is underway. To participate, visit our Consultations on the 4th review page and respond to the survey. Consultations are open until January 8, 2025.

To determine the scope of the 4th review, initial steps involved completing literary reviews, jurisdictional and legal scans for a number of topics in the context of automation and artificial intelligence.  

Next steps include:

  • Drafting recommendations – complete
  • Targeted consultation and feedback analysis – underway
  • Governance and approvals – winter 2025
  • Publication – late spring 2025

Guidance

Treasury Board of Canada Secretariat (TBS) has developed guidance to support compliance with the directive:

  • Scope – to support departments in understanding the systems and uses that fall under the purview of the directive
  • Peer review – to support departments commissioning a review and individuals undertaking the review
  • Section 4.5 of the Guideline on Service and Digital provides additional guidance on the responsible and ethical use of automated decision systems
  • Algorithmic Impact Assessment Guide – to support departments in completing the Algorithmic Impact Assessment and improve the consistency and quality of questionnaire responses

Upcoming:

  • Reporting – to support departments on the requirement of publishing information on the effectiveness and efficiency of a system in meeting program objectives

Resources

Enquiries

Please contact the TBS Responsible Data and AI team (ai-ia@tbs-sct.gc.ca) for any questions.