Canada's proposed federal budget prioritizes AI investment, initiatives

When Canadian Prime Minister Mark Carney won reelection in April, and ushered in his cabinet, he created the "Minister of Artificial Intelligence and Digital Innovation" — a post inaugurally occupied by Evan Solomon. 

In a world where the words "artificial intelligence" are sprinkled about like donut toppings, creating a minister of AI showcased that the administration prioritizes AI commercialization and utilization. Within the proposed 2025 federal budget, which passed in the House of Commons 17 Nov. and is now seeking Royal Assent, the "how" has further crystalized into something that can be transformative if implemented well, or bluster if the mark is not met. 

The budget aims to catalyze more investment in AI infrastructure, and create AI tools and initiatives that will save taxpayers money and improve service delivery. 

AIinfrastructure is a priority

For external AI goals, Carney ran on infrastructure projects and nation-building, and it looks like data centers will be a part of that. Carney's budget uses investment tax credits, including retroactive inclusion of past tax credits, to attract investments in projects like data centers and housing. He views public-private partnerships as one catalyst in his nation-building projects, with CAD280 billion over five years that will be used to support third parties seeking to build. 

In addition, the budget states the administration intends to let the Canada Infrastructure Bank invest in AI infrastructure projects. 

Carney wants AI to make government more efficient

The budget states it will transform how government works and that a new Office of Digital Transformation "will lead the adoption of AI and other new technologies across government. 

"This, coupled with near-term procurement of made-in-Canada sovereign AI tools for the public service, means not just a more efficient government, but also a generational opportunity for our domestic innovators."

The budget specifically states how some government entities will use AI, and their hopeful results. For instance, the "Department of Justice will integrate AI, advanced analytics and automation tools to streamline routine tasks, enhance decision-making and free employees to focus on higher-value strategic work."  

The current administration wants AI in government, to make it quicker and more cost-effective, and to have government AI tools developed both in-house and outside. To follow through with these ambitions, Shared Services Canada — working with the private sector, the Department of National Defence and the Communications Security Establishment — is developing an AI tool that can be used by public services across the federal government.

AI to help save 15% of expenditures in three years

The budget cites AI as one tool among many that will help "meet up to 15 per cent in savings targets over three years" for a few entities, such as the Department of Justice, Public Services and Procurement Canada, the Department of Fisheries and Oceans, and Natural Resources Canada.

Citing efficiencies from AI as one way to cut expenditures will ultimately depend on implementation and execution. If AI can allow a more productive public sector that improves the life of workers and citizens, terrific. But it may be hasty to bank on efficiency cuts before or in early implementation of AI. To varying extents, the budget describes the specific aims of how AI will be used by specific entities.

Stats Canada — The how of AI

To look at how the private and public sectors use AI, Statistics Canada will be given CAD25 million over six years to implement the "Artificial Intelligence and Technology Measurement Program" that aims to gain insights on how AI is being used in Canada, and how it is affecting workers, businesses and broader society. 

A macro analysis of how AI is affecting society is useful. Wading through what is AI hype, versus actual substantive effects and harms in an information ecosphere that prioritizes hyperbole can possess utility and inform policymaking. A more granular view of the use cases that both help and harm can also create useful what-to-do and what-NOT-to-do use cases. 

Department of Justice — AI and administrative law

The Department of Justice is set to "integrate AI, advanced analytics and automation tools to streamline routine tasks, enhance decision-making and free employees to focus on higher-value strategic work." It's a bit vague, but the "enhance decision-making" part is of particular interest. If that decision-making is referring to formal decisions made in tribunals and other government entities, issues regarding hallmarks of administrative law, such as bias and delegation may arise. 

Further integrating AI into federal administrative law will likely result in updates to the Directive on Automated Decision-Making, and some interesting new case law from the courts.

Canada Revenue Agency — AI amid embattled customer service

The budget states that "AI and process automation will be leveraged" at the CRA, "to transform technologies, data and analytics systems for compliance and collection activities." An effective tax agency can yield terrific return on investment for governments, the CRA was scorched after a recent auditor general report for difficult and inaccurate customer service.  

In 2024-25, only 18% of calls were answered within target, and less than 5% of calls were answered within 15 minutes. Of test calls by the auditor general, only 17% were accurate regarding individual taxes, 56% regarding benefits, and 54% regarding business taxes. Most relevant to Carney's AI aspirations is that the CRA's chatbot was accurate about the information it provided just one-third of the time.

The CRA is in desperate need of improvement, but for the budget's AI promises — relating to both the CRA and the public service as a whole — to come to fruition, it must do much better than a chatbot that's less accurate than a coin flip.

Only time will tell

Budgets show where money is to be allocated, providing a paper trail that can foretell ambitious success or miserable failure. There's a lot of money involved here, and only three years to see if this check will come through, or bounce.

Warren Urquhart is an associate at Murdoch Legal with a multi-jurisdictional practice serving Canadian and American clients.