Government of Canada AI Register (Minimum Viable Product)
Canada · Framework · Treasury Board of Canada Secretariat
Our summary
A public database documenting artificial intelligence systems deployed across Canadian federal government departments and agencies. The early version assembles data from multiple institutional sources including impact assessments, freedom of information submissions, and parliamentary documentation. The registry will evolve based on stakeholder input before being replaced by a more comprehensive successor with regular updates.
Written by the Index in our own words. It is not legal advice; the official text is the authority.
Inside this policy
Every section of this policy in order. Our notes summarise each one in plain language and do not quote the official text, which is the authority.
Drafted by the Index An automated summary of the official text, not yet reviewed by a person.Reviewed Checked by a person against the official text.
8 sections
Jump to a chapter
Insect collection label transcription project
The Government of Canada operates or develops AI systems across multiple departments for diverse uses: legal research, customer service chatbots, document processing, crop and disease identification, border security, radiation detection, translation, content generation, facial recognition for identity verification and travel processing, anomaly detection in financial systems, and code analysis. Systems range from production (since 1994 or 2009 for some, 2023-2024 for others) to development stage, with applications for government employees and the public, and success metrics including at least 15% of IT calls handled, at least 50% reduction in administrative or analysis time, at least 70% reduction in manual document work, and at least 90-95% accuracy for nutritional data extraction.
Canada School of Public Service AI systems
This register documents 32 AI systems deployed, in production, in development, or piloted across Canadian government departments including the Canada School of Public Service, Canadian Air Transport Security Authority, Canadian Heritage, Canadian Space Agency, Correctional Service of Canada, Department of Justice Canada, and Employment and Social Development Canada. The systems support diverse functions including conversational chatbots for recommendations and inquiries, sentiment analysis, security threat detection, content generation, legal research and document review, benefits administration and eligibility assessment, and specialized tasks such as fatigue detection, explosive detection, and disability case file organization. Systems are used by government employees, the public, or both, and include vendor-supplied and in-house developed tools using technologies such as generative AI, machine learning, natural language processing, computer vision, and retrieval-augmented generation.
Data entry for grants and contributions
The Government of Canada AI Register lists 33 AI projects across multiple departments, most in development or production for government employee use, including systems that automate administrative tasks (robotic process automation, document classification, filtering and summarization), support service delivery (social insurance processing, benefit calculations, export permits), enhance forecasting and analysis (climate prediction, oil spill detection, oceanographic monitoring), and assist information management and decision-making (complaint analysis, news topic detection, project documentation classification, financial risk monitoring). Some retired projects from Global Affairs Canada (2019-2023) used natural language processing and machine learning for foreign policy analysis and financial auditing but were discontinued due to accuracy issues, integration challenges, privacy concerns, or insufficient value. A pilot by Environment and Climate Change Canada is testing Microsoft M365 Copilot for drafting documents and emails, with deployment decisions pending procurement completion.
Enhanced Reporting Capabilities for International Assistance Grants and Contributions
Global Affairs Canada operates or develops AI tools for policy assessment, document analysis, translation, media monitoring, search and analysis, drafting briefing notes, data verification, routine inquiries, employee chatbots, internal knowledge platforms, and media summarization. Other Canadian government departments deploy AI for scientific literature review, pesticide assessment, product information collection, financial forecasting, bridge monitoring, immigration application processing including eligibility determinations and fraud detection, passport triage, administrative tasks, and environmental impact assessment, with tools supporting staff efficiency and decision-making while maintaining human oversight in final determinations.
M365 Copilot pilot
The Government of Canada has registered 33 AI applications across multiple departments, including conversational tools for employees (LibreChat, Radia, RadiaPRO, Copilot pilots), public-facing services (trademark and design analysis, business finders, Ombot), and internal efficiency systems (email summarization, code generation, speech writing, document processing, data quality monitoring). Applications process unclassified or sensitive information up to Protected B level, operate in cloud environments, and aim to reduce manual workload, improve processing speed, and enhance decision-making, while maintaining human oversight and data security standards.
AI textual analysis tools
Canada's government uses AI systems for document management and cultural preservation (Library and Archives Canada classified 25,000 files and transcribed 117,000 newspaper records and 6 million historical pages with error rates below 7 percent), research support (National Research Council tools improve nutrition prediction, traffic planning, molecular discovery, and disease monitoring), infrastructure and scientific applications (unmanned aerial system detection, bridge inspections, hydrogen leak monitoring, harbour ice forecasting, drone operations, and driver safety prediction), translation (neural machine translation for Parliament), and environmental analysis (geospatial services classification, energy optimization, flood susceptibility mapping, and satellite-based feature extraction). Most systems are in production or deployed; some remain in development as of 2025.
Historical Imagery Processing and LiDAR Classification
Natural Resources Canada operates AI systems for geospatial data processing, earth observation, materials science, industrial processes, and semantic search on public portals. Other federal departments use AI for research support, document automation, parliamentary transcription, software development, language models, translation, procurement processing, IT support, fraud detection, and policy assistance. Law enforcement uses AI for child exploitation identification, video search, transcription, facial matching (disabled), evidence categorization, and cybercrime coordination.
Monitoring Canadian waterways for obstructions
The Government of Canada registers multiple AI systems across departments, including machine learning for cybercrime severity assessment, generative AI translation and audio transcription tools, large language models for government employee productivity and public-facing chatbots, recommendation systems for survey and manufacturing data analysis, automated systems for cargo and vehicle monitoring, and AI tools for document processing, translation, and fraud detection. Systems range from production deployments and proofs of concept to development and retired systems, serving government employees and the public, with applications including workflow automation, sentiment analysis, risk assessment, regulatory analysis, and handwriting recognition for accessibility and efficiency. The register covers systems by the RCMP, Statistics Canada, Transport Canada, Treasury Board, Veterans Affairs Canada, Shared Services Canada, and other departments, many incorporating protected data handling, human review requirements, or bilingual capabilities.
Controls
What this policy requires, forbids, grants or penalises, one control at a time and in our own words. Each control links to the official text, which is the authority. This is not legal advice.
Drafted by the Index Drafted automatically from the official text, not yet reviewed by a person.Reviewed Checked by a person against the official text.
Themes covered 7 of 35
Choose a theme to see only its controls. About the themes
Risk and scope 2 of 4
- Risk classification, not covered
- Prohibited practices, not covered
Data and privacy 0 of 4
- Data governance and quality, not covered
- Data subject rights and consent, not covered
- Cross-border data transfer, not covered
- Copyright and training data, not covered
Transparency and records 2 of 4
- Labelling of synthetic content, not covered
- Record keeping and logging, not covered
Technical safeguards 2 of 4
- Bias and non-discrimination, not covered
- Testing and evaluation, not covered
Assurance and lifecycle 0 of 4
- Conformity assessment and certification, not covered
- Registration and notification, not covered
- Incident and breach reporting, not covered
- Post-market monitoring and corrective action, not covered
Rights and protected groups 0 of 4
- Automated decision-making rights, not covered
- Children and vulnerable groups, not covered
- Biometric identification and surveillance, not covered
- Whistleblowing, complaints and redress, not covered
Roles and organisation 1 of 6
- Supply chain and provider-deployer duties, not covered
- General purpose and foundation models, not covered
- Training and AI literacy, not covered
- Public sector use and procurement, not covered
- Innovation support and sandboxes, not covered
Enforcement and sectors 0 of 5
- Enforcement, penalties and regulators, not covered
- Sector rules: health, not covered
- Sector rules: finance and insurance, not covered
- Sector rules: employment, not covered
- Sector rules: education, not covered
The full list of controls could not be loaded, so only the first 10 are shown.
Pro B AI Chatbot and Agent Interface 3 controls
- ObligationApplies to: Public authorityDrafted by the Index
Indigenous Services Canada is developing an interactive AI tool that integrates multiple models to support departmental use cases through agents, knowledge bases and limited tools for government employees.
Pro B AI Chatbot and Agent Interface - ObligationApplies to: Public authorityDrafted by the Index
Early phases of the Pro B AI Chatbot will focus on internal service data from standard sources comparable to existing solutions before incorporating more sensitive data.
Pro B AI Chatbot and Agent Interface - ObligationApplies to: Public authorityDrafted by the Index
The public authority must complete risk assessments before expanding the Pro B AI Chatbot to email, calendar and SharePoint integration and more sensitive data.
Pro B AI Chatbot and Agent Interface
AI textual analysis tools 2 controls
- ObligationApplies to: Public authorityDrafted by the Index
Library and Archives Canada conducted a completed pilot project to automatically categorize 25,000 digital archival files using document classification.
- ObligationApplies to: Public authorityDrafted by the Index
Library and Archives Canada operated a completed pilot project that used Transkribus AI to transcribe historical records from 1925-1940, recognizing handwritten and printed text for approximately 117,000 Canadian newspaper records with minimal error rates.
Historical Imagery Processing and LiDAR Classification 2 controls
- ObligationApplies to: Public authorityDrafted by the Index
Natural Resources Canada operates an AI system that automatically processes and geo-references historical air photos through alignment, super-resolution and colorization for temporal change analysis.
- ObligationApplies to: Public authorityDrafted by the Index
Natural Resources Canada operates an AI system that classifies buildings, ground and water features from LiDAR data by providing trained models and documentation to production teams.
RCMP AI applications for cybercrime and translation 3 controls
- ObligationApplies to: Public authorityDrafted by the Index
The RCMP applies machine learning and natural language processing to optimize resource allocation and enhance report triage and priority assessment for cybercrime and fraud incidents according to defined business rules.
RCMP AI applications for cybercrime and translation - Applies to: Public authorityDrafted by the Index
The RCMP uses vendor-supplied DeepL Pro for automatic translation of unclassified information to support official language requirements.
RCMP AI applications for cybercrime and translation - ObligationApplies to: Public authorityDrafted by the Index
The RCMP is developing a generative AI tool called Draft One that creates draft narratives from body-worn camera audio for officers to review and finalize.
RCMP AI applications for cybercrime and translation
Change history
No changes recorded since we began monitoring this policy.