Overview

On this page:

• Why an AI Strategy?
• Vision
• Principles
• Scope

Why an AI Strategy?

Artificial intelligence (AI) presents unprecedented and wide-ranging opportunities to enhance the public service and the services it offers Canadians. AI can unlock capabilities beyond human limits, opening doors to new ways of working and operating. With AI, we can create new types of services to better meet the needs of those we serve and improve the quality and efficiency of services already offered.

AI can streamline or automate routine tasks for public servants, freeing them to focus on more complex and critical work. It can increase the public service’s efficiency, effectiveness and productivity, maximizing its value to Canadians. It can improve the speed and scale of data and information analysis far beyond what was once possible, leading to faster, more informed decision making and scientific discovery. It can also create new avenues for public engagement and help us protect Canada’s interests by enhancing our ability to protect our IT and physical infrastructure.

AI is not new to the Government of Canada: departments have been using it for decades. Often developed in-house, early AI applications typically served very specific purposes and were used only by specialist staff. Over the past few years, however, AI capabilities have advanced rapidly, particularly in the field of generative AI. These capabilities are now embedded in a wide range of commercial software, making them more accessible to all public servants.

Learn more about: AI in the Government of Canada

Below are some examples of how AI is being used to improve the work of the Government of Canada.

Case processing: Immigration, Refugee and Citizenship Canada’s Advanced Analytics Solutions Centre

The Advanced Analytics Solutions Centre at Immigration, Refugees and Citizenship Canada has been using AI-based models to triage applications for temporary and permanent residence and find those which can be automatically identified as eligible. These models have been used to accelerate the processing of more than 7 million routine cases, allowing case officers to focus on more complex cases, and to strengthen program integrity by identifying fraud patterns.

Serving clients: Agriculture and Agri-Food Canada’s AgPal

 AgPal helps farmers and agri-businesses find information about over 400 federal, provincial, territorial, and municipal programs and services, along with market intelligence and research. Its generative AI tool, AgPal Chat, helps users find relevant funding and resources faster, supporting the sustainable growth and competitiveness of the sector.

Supporting public servants: The Public Services and Procurement Canada Human Capital Management AI Virtual Assistant

Human Capital Management AI Virtual Assistant is designed to support Pay Centre compensation advisors in processing pay cases from the backlog. The assistant automates routine tasks, allowing advisors to focus on complex cases and expedite resolutions. By shifting from manual to digital processes, it enhances efficiency, reduces workload, and minimizes errors.

Conducting research: Statistics Canada

Statistics Canada works with provinces and territories to collect data for the Canadian Coroner and Medical Examiner Database and organize the collected data into coherent datasets. Analysts can then assess the datasets for patterns of death over time, detecting trends to understand growing hazards to public health.

Transcribing and summarizing: Innovation, Science, Economic Development Canada’s (ISED) AI Accelerator

ISED developed a tool for its Parliamentary Affairs Unit that uses AI and open data to transcribe and summarize parliamentary committee meetings. By reducing manual notetaking, it frees Parliamentary Affairs Officers to dedicate more of their skills to analysis and interpretation, improving efficiency and employee well-being by cutting down on overtime.

Increasing productivity: Shared Services Canada (SSC)

SSC is piloting its multilingual conversational chatbot CANChat as an in-house alternative to commercial generative AI tools. CANChat can support drafting, editing, researching, summarizing, and information and data management and analysis. For greater data security and privacy, CANChat ensures that all data is safeguarded and stored in Canada, and that prompts are not used to train its AI.

Securing networks: The Canadian Centre for Cyber Security’s Assemblyline tool

The Assemblyline tool works to defend federal and critical infrastructure systems from cyber threats involves detecting patterns in vast quantities of data—something AI tools are ideally suited to. Since 2017, its Assemblyline tool has used machine learning to analyze malicious software, scanning over 1 billion files a year for over 300 Government of Canada and critical infrastructure organizations

Securing borders: Transport Canada’s Pre-load Air Cargo Targeting (PACT) program

The PACT uses AI to screen inbound air shipments before takeoff to flag those that could contain concealed explosives or other threats. The use of AI has enabled a tenfold increase in the number of shipments screened per hour and increased coverage from 6 percent to 100 percent of flights, greatly increasing safety on both passenger and cargo flights.

These advances have also greatly increased the risks of AI and the challenges of managing them. They have heightened the potential for AI to overturn traditional ways of working in the public service and have increased public scrutiny of government use of AI.  Beyond ethical and security concerns, challenges such as talent shortages, infrastructure gaps, technological sovereignty, and interdepartmental collaboration further complicate AI adoption. Existing and future AI systems must therefore be appropriately governed, with guidance, policy, and training in place to manage risk, address challenges, and uphold human rights, public trust, and national security. At the same time, there are real risks and opportunity costs if we fail to adapt to these new technologies.

The Government of Canada needs an AI strategy to ensure that its AI adoption and use:

• Is aligned with the government’s values and ethics, objectives, and mandate to serve Canadians
• Prioritizes uses that will meet the needs of and deliver the greatest benefits to public servants and those they serve
• Is developed efficiently and collaboratively with internal and external partners
• Is responsible, safe, and secure, mitigating threats, risks, and harms to people and the environment

Vision

By responsibly adopting AI, the Government of Canada ​can deliver world-class services to its clients, protect our people and interests, achieve a more innovative and efficient workplace, and accelerate scientific discovery for the benefit of all.

Principles

Human centred

We focus on the needs of those we serve and the public servants who serve them in deciding where we adopt AI and how we integrate it into our work.

Collaborative

We work together on AI adoption with Indigenous and Canadian partners, other Canadian and international jurisdictions, and our public service colleagues.

Ready

We have the data, infrastructure, tools, culture, talent, skills and policy we need for responsible, safe, secure, and successful AI adoption.

Responsible

We inform clients and public servants when and how we use AI so that they trust that our use of AI respects privacy and is justified, responsible, fair, safe, and secure.

Scope

Many different definitions for AI exist, including the definition in the Government of Canada’s Directive on Automated Decision-Making. One of the most widely accepted definitions comes from the Organisation for Economic Co-operation and Development (OECD), which defines an AI system as:

a machine-based system that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments. Different AI systems vary in their levels of autonomy and adaptiveness after deployment.

AI includes:

• Both knowledge-based systems that use a combination of domain knowledge, rules, facts, and relationships curated by human experts and machine learning systems that that can learn from data and generalize to perform tasks without explicit instructions.
• Application areas such as computer vision, natural language processing, speech recognition, intelligent decision support systems, and intelligent robotic systems.

This current strategy applies to all types of AI technologies with adaptive capabilities after initial training:

• At any stage of the AI lifecycle, from design, development, deployment, and operation to system decommissioning.
• In use for any purpose in any department or agency covered by the Policy on Service and Digital, including those used by external subcontractors. Organizations that are partially or wholly exempt from the Policy on Service and Digital, including those that fall under the national security exemption, are encouraged to comply with the AI Strategy as far as possible as a matter of good practice.
• Whether they are developed within the Government of Canada, open source, commercial off-the-shelf products, or custom vendor solutions.

The AI Strategy excludes:

• Systems that only use software-based solutions without adaptive capabilities after initial training
• The adoption of AI by organizations outside the Government of Canada.

Learn more about: How we used AI in developing this Strategy

The team responsible for the development of the AI Strategy used an approved generative AI tool (Microsoft Copilot) and some Microsoft Teams AI capabilities to support the work of its members.
AI was used during development to:

• Automatically transcribe some discussions
• Summarize and group comments and feedback
• Generate meeting invitation text
• Scan and summarize research
• Translate small pieces of text
• Draft and edit reports and discussion papers

All uses of AI were consistent with TBS policy and guidance. Any personal identifiers, such as individual or organization names, were removed before AI use. Products generated by AI were reviewed by human analysts and were labelled to indicate that AI had been used in their development.

In total, these uses of AI saved analysts approximately three weeks’ work during the development process, enabling the team to allow more time for engagement and consultation.

For more information, please see Guide on the use of generative artificial intelligence - Canada.ca