It would be nice, delightful even, if the country’s current problems were technical in nature. But technology is always second order to politics. Technological systems reflect the politics of their environment and their funding. A government intent on increasing austerity, coupled with an enthusiasm for militarization, has no capacity to execute an inspiring formulation of digital sovereignty.
The idea that a country wants to have agency over its technology systems is understandable. But let’s take care not to confuse this desire with anything resembling progressive politics. To achieve a progressive digital sovereignty requires investment, work, and power shifting beyond anything current and past governments have shown the stomach for.
The chaos of the moment will not be the last in a line of upheavals, including an evolving relationship with our southern neighbour. In this context, we should be making all possible efforts to refuse tech policy that installs more rigid and stale colonial nationalism directly into our infrastructure, creating an ever more brittle culture. What’s needed is a pluralistic and adaptive tech policy paradigm focused on localization, openness, repair, and maintenance. A paradigm that prioritizes investments in the people around the tech, not the tech itself.
This paradigm includes pragmatic acceptance of the current state—and use—of digital infrastructures in both public and private sectors. Consider three parts of this landscape: legacy infrastructures, the public sector data environment, and digital public infrastructure.
Legacy Infrastructures
Legacy infrastructures, in both the public and private sectors, are complex, deeply built out, and enmeshed. These are decades-old sets of investments in large, highly used, and hardened systems—from telecommunications to software to databases, many of which are not Canadian. These are, generally, monolithic systems, in engineering shorthand. From a technical angle, these aren’t the types of systems you change unless you must. In the public sector context, legacy systems are underfunded, taken for granted, and ill-maintained.
The idea that we must have Canadian versions of these systems is non-viable. It would cost hundreds of billions of dollars to replicate existing capacities, many of which are as banal as case management systems. Databases. These systems have been over-engineered, to our detriment. And while there is likely some value in understanding how to reorganize them in incremental ways, it would be a major work effort.
Much of our digital environment is like capital expenditures in decades past—things like highways or factories. The pieces aren’t easily movable, and it would be folly to build them twice. There is, however, work to be done to build on top of and around them—but the dependencies will remain. Interdependence is not something that we can eliminate. There are hard constraints on the marginal value of investing public monies in managing these engineering realities differently.
The public sector data environment
Next, turn to the public sector and its approach to data management. Recently, federal leaders have made announcements that signal desire to apply artificial intelligence to public service operations at all orders of government. There is, however (and thankfully), a wide gap between the realities of the public sector data environment, both federally, provincially, and at the local level, and the idea that major productivity gains are around the corner via automation.
Government-held data is subject to a range of regulations and conditions. At the operational level, there are still workflows that depend on manual extraction and copying from system to system. Beyond these technical frictions, there are also political, administrative, and legal ones—federalism, public sector culture, organizational design, and engineering constraints. These factors generally combine and conspire to stop anything from happening too quickly in public service reform. Which is a good thing.
Meanwhile, the federal government recently awarded a company named Cohere some type of access to our public data systems through a memorandum of understanding. The announcements about the endeavour are vague: improving public service operations and expanding commercial opportunities, with an emphasis on the value of supporting a Canadian company. There are numerous reasons why we can’t run any old commercial product through a public sector data environment, but whether it’s Canadian or not, we have to reorganize this discussion into one about labour rather than one about domestic innovation.
Most of the computational infrastructure in government bureaucracies manages banal tasks. Politicians’ recent enthusiasm to push public sector automation is driven by a longstanding political interest in efficiency. The question for policy here, as ever, is what, exactly, should be subject to automation, and why. One could argue that having more data scientists on hand is a better adaptive approach for long-term public benefit than any automated tooling.
Digital public infrastructure
Finally, there is the matter of digital public infrastructure. This is a useful site for more attention in our digital sovereignty conversations. There is a track of digital public infrastructure work currently underway by the federal government, one that requires immense amounts of inter-departmental coordination. The three core parts of a digital public infrastructure program, as per the 2023 G20 multilateral consensus and guiding principles on digital public infrastructure, are: digital identity, digital payments, and data exchange systems.
Digital public infrastructure requires careful development, public engagement, and ongoing investments in maintenance to ensure issues related to democratic legitimacy, consent, accessibility, and surveillance are addressed. This program must also always include parallel investments in analogue systems for equity and redundancy purposes, as well as be optional rather than mandatory in its use.
Investment in digital public infrastructure is one approach to ensure that the increased focus on digital sovereignty does more than enrich a small number of Canadian tech firms—a handful of which are well-situated to receive billions of dollars in public money.
Digital public infrastructure helps create the conditions for more place-based innovation policy and effort, outcomes that build on the federal government’s commitment to deliver universal high-speed internet access by 2030. This approach offers an alternative to the extractive shape of data centre placement politics—a political economy we should be running away from, not towards. It also encourages us to use best-of-class technology, as per international standards.
A digital industrial strategy
The real world of digital sovereignty should be about investing in people over products. Digital anything is secondary to primary responsibilities of reconciliation, repair, and equity in service delivery and procurement.
The path we’re on instead is techno-solutionism: the cheapest, straightest line governments can choose towards efficiency and savings, and the most impoverished innovation strategy possible for a young country with every opportunity to do differently.
