Canadian companies cannot afford to wait for homegrown artificial intelligence, argues Neil Cawse in a recent opinion piece. While the federal government's new AI strategy, AI for All, is welcome, Cawse emphasizes that rather than owning every server, Canada should focus on controlling the data that runs on them.
Federal AI Strategy: Ambitious but Needs Focus
The federal government recently released AI for All, a national AI strategy backed by $2.3 billion in funding. The plan aims to increase business AI adoption from 12 percent to 60 percent by 2034 and includes a clear commitment to data sovereignty. For those working at the intersection of technology and physical industry, there is much to appreciate.
Transportation is named as a priority sector, and Cawse agrees this is the right call. The movement of goods and people underpins every other sector, including health, energy, agriculture, and manufacturing. AI is already transforming fleet operations, driver safety, and environmental footprint measurement. Getting transportation right will have compounding benefits across the entire economy.
Avoid Wasting Money on Building AI from Scratch
Cawse warns against wasting money developing proprietary AI models or constructing data centers from scratch. The government's priority should be to keep options open, avoid committing to any single vendor or technology, and instead incentivize Canadian companies to leverage existing AI to improve productivity.
The big tech hyperscalers—Google LLC, Microsoft Corp., Amazon.com Inc., and Meta Platforms Inc.—are expected to spend more than US$700 billion on AI infrastructure in 2026. In contrast, Canada's entire AI strategy budget is $2.3 billion over five years. Trying to compete on infrastructure is unrealistic. Instead, Canada should focus on controlling data: where it lives, who can access it, and the legal framework surrounding it.
Strengthening Implementation with Safety Standards
The strategy can be improved in implementation. Voluntary certification for AI is a reasonable starting point for many applications, but transportation involves AI in safety-critical environments like braking decisions, driver monitoring, and hazard detection on highways. A minimum standard for safety, data quality, and accountability is essential for responsible deployment.
Provinces are already moving in this direction. British Columbia recently passed legislation mandating cameras in commercial trucks. The United States will soon require new heavy trucks to be equipped with automatic emergency braking technology that relies on forward-facing AI video cameras. Federal leadership on responsible AI in these environments would provide industry clarity and public confidence.
Connected Vehicle Data Needs Clear Policy
Canada also needs a clearer position on connected vehicle data. As vehicles become more AI-enabled, there is a real risk that data will become fragmented across proprietary systems, rendering it inaccessible to the businesses that generate it and incompatible across platforms. Establishing interoperability standards early, before the ecosystem hardens, is the kind of infrastructure decision that shapes an industry for decades.
