AI Content Fanning the Flames of Misinformation Online
By CityNews Montreal
Artificial intelligence is no longer a distant, abstract force in the digital world; it has become a powerful engine for creating and accelerating misinformation, producing convincing fake images, audio, and video at a scale never seen before. From the offices of Montreal’s thriving tech sector to the everyday social media feeds of Canadians, the internet is increasingly polluted with synthetic content designed to deceive, manipulate, and polarize. Experts interviewed by CityNews Montreal warn that the same AI technologies being developed and celebrated in Quebec’s world-renowned research hubs are being weaponized by bad actors to erode public trust. Deepfakes of political leaders, chatbots spreading fabricated news stories, and algorithmically generated images of events that never happened are no longer science fiction. They are here, and they are being shared thousands of times before anyone can properly assess their authenticity. The result is an information ecosystem where truth and fiction become almost impossible to separate, and the consequences are becoming visible in election campaigns, public health messaging, and even personal relationships. The problem is not simply that false content exists; it is that AI has made creating false content so cheap, so fast, and so easy that misinformation has become a flood rather than a trickle.
One of the most alarming developments in recent months has been the use of AI-generated content in political discourse, both in Canada and abroad. In the lead-up to a major U.S. primary election, a fake robocall mimicking President Joe Biden’s voice urged voters not to participate, causing immediate chaos and a wave of investigative stories. Closer to home, Canadian election officials have expressed growing concern that synthetic media could be used to interfere with federal and provincial votes, including during future campaigns in Quebec. AI-generated videos have circulated showing political figures saying things they never said, and manipulated photographs have placed candidates in compromising or absurd situations. While political misinformation is as old as journalism itself, AI has dramatically lowered the barrier to entry. In previous election cycles, creating a convincing fake audio clip or video required technical expertise and significant resources. Now, free or inexpensive AI tools allow anyone, from foreign state actors to local internet trolls, to generate high-quality disinformation in minutes. Political operatives and independent influencers alike can use AI to micro-target voters with personalized falsehoods, drawing on data collected from social media platforms. The emotional impact of seeing a candidate appear to say something outrageous is far stronger than reading a text-based claim, making synthetic media especially dangerous for democracy.
Beyond politics, AI-generated misinformation is increasingly seeping into public health and climate conversations, where it can cause real physical harm. During recent health emergencies, medical misinformation generated by AI has spread faster than official updates, with fabricated advice about treatments, vaccines, and prevention methods circulating widely on platforms like Facebook, X, and TikTok. Montreal’s public health authorities have had to spend valuable time debunking AI-created images and false narratives related to outbreaks and hospital capacity. The challenge is that AI algorithms are designed to maximize engagement, and sensationalist or frightening content tends to attract more clicks, shares, and reactions than measured, accurate information. This means that the platforms themselves, by prioritizing algorithmic engagement, often amplify the very AI-generated content that experts are trying to correct. In the climate sphere, fake images of natural disasters, manipulated maps of future flooding, and AI-generated quotes from scientists have further polarized an already contentious public debate. The same tools that allow artists and filmmakers to create stunning visual effects are also being used to cast doubt on established science and to manufacture false controversies. In Canada, where wildfires and severe weather have become increasingly common, AI-generated imagery of disasters has been shown by researchers to create confusion about the scale and cause of events, undermining public confidence in emergency alerts and environmental reporting.
The mechanics behind this new wave of misinformation are rooted in the very design of generative AI models. Large language models such as ChatGPT and image generators like Midjourney are trained on enormous datasets drawn from the internet, including text and images that contain biases, conspiracies, and falsehoods. When prompted maliciously, these models can produce coherent, believable disinformation tailored to an audience’s fears, angers, and prior beliefs. This is not simply a matter of better fact-checking; it is a fundamental transformation of how propaganda operates. In the past, spreading misinformation required a level of human effort and editorial judgment, even for propagandists. Today, AI can write thousands of variants of the same false claim, adjust the tone and language to match a specific demographic, and deploy it through social media bots in a coordinated campaign. Researchers at Montreal’s Mila institute and other academic centers have documented how these attacks often target marginalized communities, using AI-generated ethnic slurs and fabricated images to deepen social divisions. The speed of the spread is also astonishing: a deepfake video can go around the world in hours, while scientific corrections or media debunks struggle to gain similar traction. This asymmetry between the speed of falsehood and the speed of truth is one of the most troubling features of the current moment, and it has pushed journalists and civil society groups to call for a more aggressive response from both technology companies and governments.
Governments and platforms are beginning to react, but their efforts are uneven and often too slow. Some major tech companies have announced voluntary policies requiring labels on AI-generated content, and a number of open-source initiatives have proposed digital watermarks that can help verify whether a piece of media was created by a machine. However, these measures are easily circumvented, especially because many powerful AI tools are available in open-source form, allowing malicious actors to remove watermarks or simply generate content using custom systems. In Canada, federal lawmakers have debated legislation aimed at regulating artificial intelligence, but the process has moved slowly, and experts worry that any eventual law will be outdated by the time it is passed. Quebec has also taken some steps, with local politicians discussing the need for greater education and digital literacy, but concrete action remains limited. Meanwhile, journalists and fact-checking organizations are struggling to keep up with the volume of AI-generated content, despite efforts to establish cross-platform verification networks. Montreal’s journalism schools and media outlets have started integrating AI literacy into their training, and some newsrooms are using AI tools themselves to help detect deepfakes. Yet these responses remain reactive, and the general public continues to lack the critical digital skills needed to identify synthetic content. In interviews with CityNews Montreal, experts stressed that no single law or platform policy will solve the problem; they argue for a coordinated effort involving government regulation, corporate responsibility, educational programming, and public awareness campaigns that teach Canadians to scrutinize what they see online before sharing it.
In the end, the fight against AI-generated misinformation is ultimately a fight for trust, and no institution, algorithm, or law can fully replace the human judgment required to restore it. Montreal, as a global hub for artificial intelligence research, finds itself in an uncomfortable position: the same city that is celebrated for advancing AI is also on the front lines of observing its harmful side effects. But that concentration of talent could also be an advantage. Researchers, engineers, and entrepreneurs in Montreal have the opportunity to develop ethical AI systems that prioritize transparency, accountability, and the public good, rather than simply optimizing for engagement and profit. There are already promising projects in Quebec aimed at creating detection tools, media literacy workshops, and open-source content provenance standards. But these efforts need to be scaled rapidly and adopted by social media platforms that are often unwilling to make changes that might reduce advertising revenue. Citizens also have a role to play: pausing before sharing a provocative image, checking trusted news sources, and using reverse-image search tools can all contribute to slowing the spread of AI-generated falsehoods. The threat is real, and it is growing, but it is not inevitable that the information age becomes a disinformation age. With enough pressure from users, policymakers, and the research community, the same artificial intelligence that is currently fanning the flames of misinformation can be redirected to help put those fires out. Until then, the burden falls on a public increasingly left to navigate a digital world where seeing is no longer believing—and where every image, voice, and news story must be met with a healthy dose of skepticism.
