A recent study has sparked fresh debate over how artificial intelligence systems rely on news content, with researchers warning that Canadian journalism is being widely used without adequate attribution or financial return.
The research, conducted by McGill University’s Centre for Media, Technology and Democracy, analyzed 2,267 Canadian news articles across several leading AI platforms. The findings suggest that popular AI systems including ChatGPT, Gemini, Claude, and Grok demonstrate a strong awareness of Canadian current events, indicating that they have likely been trained on large volumes of journalistic content.
Despite this apparent reliance, the study found that these AI models failed to properly credit news sources in a significant majority of cases. When asked about events covered in their training data, the systems omitted source attribution approximately 82 percent of the time.
Even when web access was enabled and AI tools referenced specific articles, researchers observed that responses often included enough detail from original reporting to satisfy user queries potentially discouraging users from visiting the original news websites. While about half of the responses did contain at least one Canadian link, only 28 percent explicitly named a Canadian news outlet.
The report emphasized that while links can guide users back to original sources, most readers may never notice where the information originated. This, researchers argue, creates a system where journalism is consumed without recognition, weakening the connection between news producers and audiences.
According to the study, AI companies are extracting value from journalism at multiple stages—using archived content for training, generating derivative outputs without clear attribution, and delivering responses that may reduce traffic to original publishers. The researchers warned that this trend could accelerate the financial challenges already facing the news industry.
The findings were released alongside a national summit in Banff, where Canadian officials are discussing the intersection of artificial intelligence and cultural industries. During the event, Culture Minister Mark Miller acknowledged ongoing concerns about copyright, licensing frameworks, and the growing volume of AI-generated content.
Artificial Intelligence Minister Evan Solomon echoed these concerns, noting that recent consultations highlighted the need for safeguards to protect creators. He stressed that questions around data usage, ownership, and intellectual property remain unresolved, but indicated that the government is open to further dialogue.
Meanwhile, tensions between news organizations and AI developers are escalating. A coalition of major Canadian media outlets including The Canadian Press, Torstar, The Globe and Mail, Postmedia, and CBC/Radio-Canada has launched legal action against OpenAI in an Ontario court. The lawsuit alleges that the company has used copyrighted news material to train its AI systems without authorization or compensation.
As AI technology continues to evolve, the study adds urgency to calls for clearer regulations and fairer practices aimed at ensuring that the journalism underpinning these systems is both recognized and sustained.
