Following the Liberal Party’s re-election for a fourth term, Prime Minister Mark Carney stated in his mandate letter to his newly appointed cabinet: “Government itself must become much more productive by deploying AI at scale [and] by focusing on results over spending.”
In response, Evan Solomon was appointed the first federal minister of artificial intelligence and digital innovation to oversee growth in the public sector’s use of AI and to scale up the industry to meet Canada’s myriad policy needs.
As such, AI is positioned as a tool to meet Carney’s other priority – delivering more while spending less on government operations.
The government’s argument is that AI can safely and efficiently perform some of the more routine tasks of the public service, removing redundancies and bottlenecks that so often give Canadians the impression the machinery of government is beyond repair.
Thus, the order of the day is more technology, potentially fewer humans and hopefully faster services for Canadians.
Reaction has been swift.
While Carney has said that job cuts are not on the agenda, public sector unions are warning that a 15-per-cent reduction in spending on government employees, as announced as part of a federal spending review, will lead to poorer quality of service for the public.
Yet, pundits and politicians across the board agree that better implementation of technology – if done with careful attention to security – can result in a leaner bureaucracy that is more responsive.
But we have little idea about what the public thinks of this agenda. Do Canadians support the use of AI to replace service provision by humans? Is this related to how much Canadians trust (or do not trust) government or other factors such as political ideology?
Canadians split on AI in public service
In May, we surveyed a representative Canadian sample of 2,000 respondents and found that support for AI was predictably mixed.
While 42 per cent of those surveyed supported the government using artificial intelligence tools to resolve bottlenecks in service delivery, 32 per cent opposed it, with 25 per cent on the fence. This plurality in favour is surprising, given the novelty of the technology.
By contrast, we found there was more agreement on reducing the size of the public service. A full 67 per cent agreed the federal bureaucracy should be significantly reduced. Only 12 per cent disagreed.
Are these two phenomena linked and what drives them?
Those who supported reducing the size of government were much more likely (by 22 percentage points) to support the use of AI than those who didn’t. Among those who disagreed with increased use of AI, more than half opposed a reduction in the public service.
We argue here much of the government’s ability to move forward confidently with Carney’s proposed agenda on AI and the public service will come down to how much Canadians trust the government to navigate these changes at a rapid pace – and how quickly results materialize.
Trust and ideology shape support for reform
Our data suggest that the more respondents trust the federal government, the more they support the use of AI in service delivery.
On the other hand, respondents with stronger levels of trust in the federal government were less comfortable with the reduction in the size of its administrative arm than those with lower trust, by a margin of 26 percentage points. Still, even those who trust government are still broadly in favour of reducing the public service.
Of course, trust in government was not the whole story. Like many other issues in contemporary politics, political ideology played a central role.
Individuals on the political right were more likely to support both the use of AI and cutting the size of the public service. They supported government use of AI by an additional 18 percentage points over those on the left. The same was true for cutting the size of government – with a difference of 38 percentage points in support between those on the right and those on the left.
In other words, while there was a consistent pattern for both issues, that magnitude varied considerably across the political spectrum.
We also saw gender identity as an important predictor of political attitudes.
Women were 16 percentage points less likely to support AI’s use by government than men, but no more or less likely to want a reduction in the size of the public service. But the gap widened when we looked at women on the right with conservative women being less likely to support reducing the public service than their male counterparts.
Remarkably, education level seems to have no effect on support for either.
Carney’s AI agenda and the road ahead
Carney has been explicit about his enthusiasm for the AI economy and a leaner public service – both of which have the potential to contribute to efficiency and growth across Canada’s lagging economy. Indeed, the party’s promised investment of $2.5 billion in digital infrastructure has the potential to be a catalyst for this.
Our data suggest that Canadians are getting onside with this agenda as well.
Yet, accountability will be key for continued public support. Those who are skeptical about reducing the human component of government activities may become even more hesitant about AI’s use if implementation doesn’t give Canadians improved services on a quicker timeline.
Disappointing the AI enthusiasts would also be bad news. If government fails to implement AI in its routine activities, it could result in less trust in government to get ahead in the AI economy.
The public is looking for visionary leadership amid instability. The careful, purposeful use of AI to automate the more mundane tasks of government could be revolutionary.
The question is how successful the government will be in using AI to lead the transformation in public service delivery and whether Canadians will be pleased with the result.
