If you ask the average Albertan why there is so much heated rhetoric lately regarding the province’s place in Canada, you will likely get a laundry list of familiar grievances. From deep-seated frustrations over oil and gas policy to the long-standing resentment of “Laurentian elites” and the complexities of federal equalization payments, the reasons are as varied as the people themselves. However, a new Canadian startup, Cipher AI, is suggesting that the intensity of this separatist sentiment might be fueled by a much more invisible and external force. By using advanced artificial intelligence to track online narratives, they are uncovering evidence that suggests we are living through a high-stakes digital arms race where domestic political tensions are being artificially amplified by foreign actors.
Based in Edmonton and Regina, Cipher AI is a spinout from the University of Regina’s Centre for Artificial Intelligence, Data, and Conflict. Its founders—Brian McQuinn, Matthew Taylor, and Cody Buntain—view their work as a necessary defense mechanism in an era where social media allows bad actors to manipulate public discourse at lightning speed. Unlike traditional academic research, which often moves at a slower pace, Cipher is designed to be a nimble, commercial tool. It uses proprietary large language models (LLMs) to perform massive-scale pattern recognition across the internet. By identifying suspicious coordination and tracking how divisive narratives evolve across platforms, the company creates a roadmap for stakeholders to understand what is being pushed, where it originated, and who might benefit from the resulting discord.
The company is careful to emphasize that it is not in the business of censoring political debate or replacing institutional oversight. Instead, it operates on a “human-in-the-loop” model, where its findings are scrutinized by independent experts, such as those at the watchdog organization Disinfowatch. Their goal is not to silence the legitimate, democratic discussions that Albertans have about their province’s future, but to shine a light on when those discussions are being hijacked by inorganic, artificial amplification. In their view, while the separatist sentiment itself has historical roots, the way it is currently being fanned online by international entities—particularly from the United States and Russia—is a modern, manufactured phenomenon that distorts the actual consensus of the population.
One of the most concerning findings from Cipher’s research is the shift in the origin of these manipulative narratives. While Russia has historically been a primary source of covert disinformation intended to sow chaos in Western democracies, McQuinn notes a significant surge in overt, MAGA-aligned messaging originating from the United States that targets Canadian domestic issues. This presents a unique security challenge for Canada, as the country’s existing infrastructure for monitoring disinformation has frequently relied on U.S.-based operators. If the primary source of the disinformation is now coming from south of the border, there is an inherent risk that American-based firms might be hesitant to identify or challenge narratives that benefit their own political administrations. Consequently, Cipher is positioning itself as a sovereign, domestic alternative that aligns with new federal directives requiring government departments to develop robust, indigenous strategies to combat external influence.
However, the rise of AI-driven disinformation detection is not without its critics, who raise significant ethical and philosophical concerns about who holds the power to define “truth.” Clifton van der Linden, a political science professor at McMaster University, warns that these detection tools are not neutral entities; they are programmed by humans and inevitably reflect the biases and parameters set by their creators. There is a real danger in centralizing the ability to flag content, especially when it is conducted by for-profit entities. Van der Linden argues that if the profit motive or government influence begins to dictate which narratives are deemed “dangerous” or “malicious,” we risk creating a system that is, at best, opaque and, at worst, an instrument of corporate or state control that remains unaccountable to the public interest.
Ultimately, the emergence of companies like Cipher AI highlights the precarious intersection of technology, national identity, and democracy in the 21st century. While there is a consensus among experts that the status quo is unsustainable—as traditional fact-checking cannot keep pace with the sheer volume of AI-generated content—many remain deeply divided on the solution. Building a resilient society will require more than just a better algorithm; it will require a transparent conversation about how we protect our democratic discourse from manipulation without sacrificing the very freedom of speech that define our institutions. As Cipher continues its work, the challenge will be ensuring that in our rush to build a shield against foreign interference, we do not accidentally build a wall that restricts our own capacity for healthy, authentic debate.