Machine Learning
The Crisis of Canadian Democracy
Andrew Coyne
Sutherland House, 266 pp.
The world is full of analysts, or at least it has been up until now. Let us consider what is about to become of them.
Analysts are people whose intellectual aptitude is keeping a close watch on a complex flow of events and information, and who then write up reports to help make sense of it all. If you work for an investment house, a law firm, a government department, in the officer ranks of the military or the management echelons of any corporation, you have either read or written an analyst’s report. A university is nothing but analysts endlessly writing reports to one another, from the undergrad whose paper is marked by the graduate student, to the doctoral student whose thesis is being examined, to the professor’s paper under peer review.
Analysts are brokers between knowing what’s happening and deciding what do about it. Unlike propaganda, analysis isn’t supposed to do our thinking for us. It’s supposed to make us think. Right now, the whispered selling point of generative AI is that the Large Language Models can fulfill this function, the way the desktop computer killed the typing pool. We won’t need analysts anymore.
Andrew Coyne is one of Canada’s preeminent public political analysts, and we are lucky to have him. For those who don’t follow politics and may not know him, he’s like one of the judges on Canada’s Got Talent, but for parliamentary democracy and public policy. His métier is a form of political scepticism, like the Senate but with more people paying attention to him.
I’ve been following his work for 30 years as he has opined in the Globe and Mail, the National Post, the CBC, back to the Globe and Mail, etc. He’s thoughtful, perceptive, astute, informative, quick-witted, and he writes really well. Sometimes he’s persuasive and sometimes he isn’t, but he almost always jogs you with some perception or point of argument you would not have thought of. He can get worked up but he’s never inflammatory. Quite the opposite, his rationality helps to keep things calm. He’s everything we want in a political columnist. Is he also the last of his kind?
Could an LLM trained on the 30-year corpus of Coyne’s writing, speeches, radio and television appearances, anticipate his reaction to a political turn of events sufficiently well that it could write the column without him and be close enough you couldn’t be sure if he wrote it or not? Would it matter?
What might an AI version of Andrew Coyne look like, as a way to make money? Let’s say you just want Coyne-like bursts of opinion on things. Is it theft if an AI company memorizes the entirety of Coyne’s digital record in order to spit out imitations of him without compensating him? Or would it be better if Coyne endorsed an AI franchise that generated on-demand political analysis in his name, if not by his pen, like a George Foreman Grill for politicos in a hurry?
I, a mere human, am not a Large Language Model, but having followed Coyne over the years I know something of his tics and tells. Don’t get him started on public support for news journalism, for example. For him, any public money that goes toward news reporting means someone made a decision about who would get the money, hence commissars are in charge of how people will be politically informed, not the market. That’s Coyne’s big tell. He’s an absolutist.
I figure it comes from his education in economics. To this day he thinks economics trumps other ways of understanding current affairs and human history. Economists believe there are laws of trade and banking and the value of money that can no more be defied than the law of gravity. This is their mistake. It makes them absolutists, while those of us untouched by degrees in economics can see humankind as a living thing that creates its own laws, and then defies them.
Which brings us to The Crisis of Canadian Democracy, Coyne’s analysis of the perilous state of present-day Canada. He has spent a career chronicling our partisan arena and the constitutional framework that is supposed to impose order on the whole ill-behaved mess. This is his summation of where we stand. An absolutist looks at Canadian politics, with its referendums and its shaky constitution that Quebec didn’t sign, its built-in friction between the provinces and the feds, and he sees only catastrophe ahead. There is no happy, prosperous Canada of tomorrow unless we repair the political structure that got us this far.
The complaint, not a new one, is that our democratic system is nowhere near democratic. It is a distorted, manipulated, untrue register of political will. The first-past-the post system. The unfair weighting of votes from different ridings. The Senate. Party membership drives, leadership contests, and riding nomination procedures. The meaninglessness of being an MP, the end of government by cabinet, the rise of political consultants, and the consolidation of power in the Prime Minister’s Office. It’s all a sham democracy, really. We get the governments we ill-deserve, Coyne argues, because we are running 21st century politics through 19th century plumbing.
What he fears and foretells is The Fall of Canada. Not through annexation by some demented megalomaniac, but because of our own self-worsening. We take our stability for granted when it’s not guaranteed at all, and it’s the faith in ourselves that will be our undoing. “Part of the crisis,” he says, exasperated as all get out, “is that so many people do not see it as a crisis.” If we don’t fix the unfixable, either the country sleepwalks into constitutional collapse and dissolution, or pressure points in the electorate will tear the place apart.
In this way, Coyne helps to keep the place together.
Just as Churchill said that truth, in time of war, must be protected by a bodyguard of lies, so a true and working democracy must be accompanied by a halo of complaint from serious people about how it’s not as democratic as it wants you to believe. That’s how you know it’s a free society. Coyne has long been a resolute member of this bodyguard of complaint.
He has some suggestions about how we might un-worsen things. Some are worth considering and some, I have to say, teeter on the cockamamie. He has a whole theory about how the incivility in the Canadian House of Commons could be undone by reconfiguring the chamber in which they meet. He believes it is not sufficiently cramped, and he doesn’t like the fact that our MPs have desks, which back in the day they used to thump. He prefers the UK model, where MPs crowd into tiered benches, from a time before lumbar support had been invented.
He nods quite a lot to the UK parliamentary system as superior to ours. But measured by outcomes, is the UK a better governed country than Canada? No, it is not. (Ahem, Brexit.)
What Coyne admires about the US system is that its constitution is a binding document in which the division of powers is written down and fixed, guaranteeing inviolable checks on arbitrary and unlawful executive privilege. Unlike our troublesome Canadian constitution, an unbinding document with the notwithstanding clause rattling around inside it like a weight in loaded dice.
The book clearly went to press before Trump reassumed office in 2025 and made arbitrary, unlawful executive privilege his signature move. Turns out a sternly worded constitution is not by itself enough to keep a country from going off the rails.
Coyne is right to raise his voice about the crisis of Canada. But the extinction-level political disaster he’s so worried about hasn’t happened over the span of this country’s ungainly, unworkable existence. Which means that Coyne could be right tomorrow but so far has been wrong for the past 158 years.
Perhaps the incompatibility of Canada’s working parts is not a constitutional deathtrap but a thing that must be negotiated forever, a reality where nothing is certain, everything is contingent, and decisions are deferred. This is no country for absolutists.
And how will we know about our politics in the Canada of tomorrow? Who will inform us, how will they do so, and for whom will they work? An AI version of Andrew Coyne is conceivable, though an AI version of Kevin O’Leary would probably make more money. But the real money, the tetra-billions of money, will come from the transformative machine takeover of the analysis function in every corporation and government ministry in the world.
Isn’t that right, Minister Solomon?
eg The Hill Times December 4, 2025