Imagine you are scrolling through Facebook and come across an article titled “Ottawa Betrayed the West Again—Who’s Surprised?” It comes from a publication you’ve never heard of, but the layout looks like a typical small-town Alberta newspaper. There’s a weather box, a “latest news” sidebar, and a byline photo of a smiling woman whose name seems strangely hard to remember. The article is full of angry references to equalization payments, federal overreach, and “Eastern elites” conspiring to keep the oil industry down. You share it because it says exactly what you’ve been thinking. Then someone comments: “That’s a fake website.” You check the site’s About page, and the wording is tangled. The contact email is a Gmail address. The office photo is an image you vaguely recognize from a corporate website. That’s the world described in CTV News’s “Deception Decoded” investigation: a sprawling network of fake local news sites designed to weaponize Alberta separatism, and now those sites are being consumed by AI chatbots, which are repeating their disinformation as if it were mainstream fact. The site might even include a fabricated poll, a quote from a “University of Calgary political scientist” whose name produces no results, and a social media post supposedly from a local mayor. The design is clean. The headlines are urgent. The comments are filled with angry accounts that appear to be bots. It is a complete media ecosystem built from nothing, and its entire mission is to make you believe that Alberta is on the verge of abandoning Canada. According to the investigation, these fake sites are not isolated pranks; they are part of a coordinated content network that has found a powerful new ally in AI chatbots. When an AI assistant is asked about separatism, it can draw on these falsehoods, sanitize them with a neutral tone, and present them to users as facts. The result is a disinformation loop that launders nonsense through machines and feeds it back to humans.
The fake sites didn’t create Alberta separatism out of thin air. Western alienation is a real, decades-old feeling in Canada, with roots in resource politics, equalization formulas, and the sense that Ottawa is permanently out of touch with the Prairies. The people behind these websites understand that, and they use it like a key. They know the emotional vocabulary of Albertans: Ottawa, taxes, pipelines, elites, oil, betrayal. So they publish stories that feel like a validation of everything a frustrated reader already believes. One article might claim that “federal officials secretly plan to cut Alberta’s energy industry down to half its size.” Another might quote an entirely invented “senior advisor” at the Prime Minister’s Office saying that Alberta is “just a dollar sign.” The fake outlets give themselves harmless names, something like “The Alberta Sentinel” or “Western Freedom Examiner,” and they fill their pages with generic photos of flag-waving crowds, oil pumpjacks, and frowning politicians. They don’t always make obviously absurd claims. The most dangerous content is the half-true material: a real policy debate, a genuine conflict between Alberta and Ottawa, then a sharp turn into conspiracy, exaggeration, or invented numbers. This is what makes the fake sites so effective. They borrow the language of legitimate journalism, and they wrap their propaganda in a familiar Canadian story. A reader in Medicine Hat or Grande Prairie might not be looking for a new country; they might just be looking for information that makes sense of their struggles. These sites deliver that information with undeniable urgency, and the reader feels heard. The problem is that the article is not reporting; it is targeting. It is designed to convert frustration into outrage, and outrage into support for separation.
The most disturbing part of the investigation is the way fake content has started feeding real artificial intelligence. Chatbots like ChatGPT and Google’s Gemini are trained on enormous amounts of text scraped from the open web. That includes thousands of low-quality, AI-generated, and entirely fake news articles. When a person asks a chatbot a question like “Why do some Albertans want to separate from Canada?” or “Is Alberta treated unfairly by the federal government?” the chatbot doesn’t consult an encyclopedia; it predicts the most plausible answer from its training data. If enough fake sites repeat the same false claims, those claims become statistically normal in the data. The chatbot will then calmly say something like, “Many Albertans feel that the province’s oil wealth is unfairly taken by Ottawa, and some separatist groups believe the federation is no longer viable.” That statement might sound reasonable, especially because the chatbot says it in a polite, confident tone. But it could be entirely based on disinformation. The investigation showed how this process gives fake websites a kind of invisible legitimacy. A person who would never trust a strange Facebook page might be willing to trust an AI answer. After all, wasn’t the machine trained on millions of sources? That’s exactly the illusion. The machine has no idea what is true. It only knows what has been written, repeated, and circulated. The fake sites are systematically polluting the wells that AI drinks from. And once a lie appears in an AI output, it gets shared on Reddit, copied into newsletters, repeated on podcasts, and then picked up by the next generation of AI training data. It becomes a closed loop where machines are whispering to each other, and ordinary people are listening in and believing them.
So who is actually behind all this? That is the question at the heart of “Deception Decoded.” And the answer is not a simple one. The public faces of the fake sites are usually fictional people. They have stock photo headshots, vague biographies, and names that sound plausible but turn out to be invented. The websites themselves are registered with privacy shields, so the real owners are hidden. Payments for domain names and hosting are sometimes made with cryptocurrency or prepaid cards. The server addresses are scattered across the world, making takedowns and legal challenges harder. But investigators found fingerprints throughout the network. Several sites share the same clickbait templates, the same tracking codes, the same unfamiliar sentence patterns, even the same errors in formatting. Some of the sites are connected to a handful of known “content farms” that can produce hundreds of articles per day using AI tools and low-wage freelance writers. Others appear to be side projects of political operatives who want to influence the conversation without revealing their involvement. The investigation does not offer a single dramatic name, but it paints a clear portrait: a shadowy, distributed operation run by opportunists and ideologues who have learned that fake news is a cheap and effective way to monetize anger. Some of them are probably motivated by money; fake news attracts clicks, clicks attract ads, and political rage is one of the most profitable commodities on the internet. Others are motivated by a genuine desire to weaken Canada. And some may simply be experimenting with AI, testing how far they can push the technology. Whatever their exact motives, the result is the same: they are using technology to exploit democratic vulnerability.
The real damage of all this goes far beyond the strange websites themselves. It changes the public conversation in subtle but serious ways. Political leaders and commentators rely on media coverage to understand what ordinary people are thinking. When fake stories about Alberta separatism circulate enough, some politicians will start talking about them as if they were real events. A blogger might cite a fake poll, a radio host might quote a fake news article, and a political candidate might say, “People across the West are fed up and talking about separation.” That phrase then gets picked up by another fake site, and soon the idea seems much more mainstream than it actually is. This is what researchers call citation laundering: a falsehood is manufactured, repeated, reused, and eventually treated as a common fact. The impact is not just that some people believe in separation; it is that everyone else starts to think the country is falling apart. That creates fear and anger on all sides. Genuine Alberta separatists are also harmed, because their real political opinions become associated with anonymous bots and foreign-looking conspiracy networks. The public becomes more cynical, more suspicious of all media, and less capable of having a productive conversation about the future of the federation. In an age of misinformation, the truth is fragile. The Canadian political system depends on a shared sense of reality. If fake websites and AI chatbots can dissolve that sense, they can undermine democracy itself without ever hiring a single political candidate or casting a single vote.
The good news is that ordinary people can respond to this threat with a basic but powerful habit: checking the source. Before sharing an article about Alberta separatism, or any emotional political story, ask a few questions. Is there an author with a real publication history? Does the website have a corrections policy, a phone number, and a physical address? Are other legitimate news outlets reporting the same thing? Search the exact headline in quotes. If the only matches are the fake site itself and some angry social media posts, that is a warning sign. It also helps to slow down emotionally. Disinformation works by making you feel something before you think. Outrage, indignation, and fear are all shortcuts past your judgment. When a headline makes your blood boil, that is the exact moment to pause and ask who wants you to feel that way. The investigation also calls on technology companies to do more. AI companies should be transparent about their training data and should actively filter known disinformation sites. Search engines should lower the ranking of low-quality content farms. Social media platforms should hire more fact-checkers, label AI-generated content, and stop rewarding rage-driven shares. None of this will be perfect, but it can make a difference. The people behind these fake sites are not geniuses. They are simply exploiting a system that cares more about engagement than truth. The antidote is not cynicism; it is healthy skepticism. We can still trust good journalism, and we can still have meaningful debates about our country. We just need to stop letting anonymous machines decide which reality we live in. The next time you see a story that feels too perfect in its anger, take a breath. Look behind the website. And remember that the person who created it probably isn’t in Alberta, and probably isn’t trying to help you.