A panel of industry professionals take on generative AI in Canadian newsrooms
“When we talk about generative AI,” says Cody Gault, “there is fear that it’s going to take newsroom jobs.” Gault, product manager at the Toronto Star, saying what was on everyone’s mind on a panel on the future of AI in newsrooms to an audience of working journalists, retired professionals, students, and professors.
Gault was joined on March 11, 2025 at Toronto Metropolitan University’s Rogers Communication Centre by three other industry experts on AI in Canadian journalism. Hosted by the Review of Journalism in The Catalyst and moderated by TMU alumnus Aloysius Wong and TMU journalism student Sandra Ingram, the four panelists dove into the ethics and implications of integrating AI in news production.
Is ChatGPT in the Newsroom?
A subset of generative AI is large language models (LLMs) such as ChatGPT and Google Gemini, which can be used for everything from software coding assistance to in-depth research. Sarah Thompson, chief commercial officer at Mantis Group with an extensive background in data evaluation and ethical data supply, describes AI as modelling applied to data to make sense of it, streamline it, and drive insight.
But how newsrooms use these tools varies. Murad Hemmadi, AI reporter for The Logic, says his use of AI in the newsroom is limited. “It might be used for research, data gathering, data cleaning, that kind of thing, but the actual words we publish will not have been generated by AI.”
Programs like Otter.ai are useful to journalists because they can generate transcriptions quickly and effectively. Gault says the tool is invaluable. “In the case of transcription, generative AI takes the worst part of the job … AI transcription is not cheating—it makes the journalism better, and makes reporters faster and more accurate.”
Mahima Singh, a data editor at The Globe and Mail, says she has been using AI in specific contexts since 2010 to assist with automation and data analysis. “It was basically something to help me make sense of a very large data set and help me automate mundane tasks and analyze things faster.”
How Is It Impacting Journalists?
Thompson believes that humans should always lead the creative process, even when integrating AI. “To me, it presents a lot of opportunities to take content and put it into different forms, but it should always be generated, first and foremost, by a human being,” she says. As for AI-generated content in the news industry, she points to existing concerns of how tools like OpenAI scrape together existing journalism and recycle it on other platforms. “LLMs or AI tools are going to syphon ad revenue from legitimate publishers,” she says. “They don’t go to The Logic, they don’t go to The Globe and Mail, they don’t go to the Toronto Star. It’s creating noise where we need to get to the human-generated side.”
The fear of what AI can mean for the industry is not only about what it can take from journalists, but what it can do when it goes unchecked. “I don’t think that the fear or the risk is that newsrooms won’t embrace AI,” says Gault. “It’s the business model that is the real threat. But if we’re zeroing in on generative AI as a threat, it’s in the form of creating deepfakes.”
Gault says identifying newsrooms that are dealing with deepfakes is the first step. “But even if we get great at that and never make that mistake, ever, we’re still competing on social media and the internet more broadly, with unscrupulous people who are more than happy to create fake footage and fake screenshots,” he says. “I do expect that I am going to be fooled at some point soon, if I haven’t already been—maybe I have—and that seems like a serious threat to journalism.”
Journalists have to reach their audiences through the overwhelming amount of fake and fabricated content that thrives online, and generative AI can help. “As an AI reporter,” says Hemmadi, “understanding how this technology works is 80 percent of my job. The other 20 percent is explaining that to other people.”
What Can’t It Do?
Though AI can help streamline many sectors of the industry such as web production, copy editing, and metadata, according to Gault, it is in these technical uses of these models that journalism can benefit most. AI can be used in these cases to fast-track dull tasks and optimize accessibility. But it’s unlikely to replace the human and personal essence of journalism.
Perhaps most anxieties over the long-term consequences and the future of AI in news can be eased by understanding the limitations of this technology. “Chatbots’ generative AI can’t go to the scene and talk to people,” says Gault. “They can’t build relationships, work with sources, or come up with novel ideas—the fundamental things journalists do.”
