A plenary discussion of the role of artificial intelligence in journalism kicked off the annual conference of the Canadian Association of Journalists June 12-13 at Carleton University.
The event’s opening panel featured four speakers and was moderated by Nikita Roy, a journalist and Harvard-trained AI futurist. Themed as the “battle for information sovereignty,” the panel consisted of Lisa Hepfner, Liberal MP for Hamilton-Mountain and a former television news journalist; Grant Ellis, executive editor of the Toronto Star; Christa Dickenson, president and CEO of CPAC; and Indigenous journalist Karyn Pugliese, APTN’s national affairs correspondent in Ottawa.
Each of the panelists expressed the importance of striking a balance between news companies’ adoption of AI to bolster their audience reach and information-gathering while avoiding risky and unethical uses of the technology. They also emphasized resisting the power of foreign AI tech companies and the scraping of news content to generate decontextualized AI search results, which may reframe and curate information incorrectly and without due credit to content creators.
Hepfner, who is also the chair of the House of Commons heritage committee, emphasized that an understanding of how AI algorithms work is essential for journalists and others grappling to adapt to the revolutionary technology.
“The education is key. We have to understand these systems,” she said. “We have to understand what they do and how they work, and then we can use them for good.”
Roy referred to generative AI as usurping the role of publisher, raising the question of the sovereignty of information produced by journalists. This is especially concerning since most users solely rely on AI-generated summaries, according to Roy, much of which is generated by content currently not sourced from local journalism due to lawsuits initiated by several Canadian media companies over AI training.
“As a citizen, I feel like I’m getting my news and information being told from companies that AI companies have licensing deals with, which are increasingly American or international companies, and I get told a story and news about Canada from a lens outside of Canada,” Roy said.
Pugliese focused her attention on the lack of nuance of algorithmic data which, when taken at face value, can give rise to preventable social issues.
“The people who program the algorithms understand the computer science, but they don’t understand the art of what they’re programming for,” she said. “I often wondered if babies being taken from Indigenous women in Canada is a result of an algorithm,” Pugliese said, citing an American report that covered Black overrepresentation in the child welfare system due to social workers relying on algorithms first to determine which families to investigate for child neglect.
On the other side of the balance was the need to innovate home-grown AI systems for use in Canadian journalism. Grant Ellis of the Toronto Star, which is currently part of a coalition of Canadian news organizations suing OpenAI, argued that not adopting homegrown AI tools would leave Canadian journalists and their readers reliant on American tech companies willing to take advantage of the emerging market.
“If we ignore it, there’s no turning back. I mean, this is a product that will mimic us, will steal from us, will distribute it on every single screen and platform possibly available,” Ellis said. “While it’s difficult to get the messaging across to journalists who rightly feel that they’re under attack by this theft from a foreign colonial power, we have to be able to use this technology, much of it developed by Canadians, frankly, to power up our journalism to fight back.”
Ellis described AI tools the Toronto Star has developed to reach specific audiences more effectively, as well as tools for beat reporters that allow them to more efficiently and rapidly peruse public documents.
For example, the Star developed an AI tool to assist investigative reporters in scraping municipal policies exercised across Ontario using the so-called “strong-mayor powers” authorized by the provincial Progressive Conservative government. The probe aimed to discover if the new powers led to less red tape for housing development (Premier Doug Ford’s principal argument for expanding mayoral power) and found that it overwhelmingly led instead to municipal staff firings and relatively inconsequential decisions about a flagpole and a pickleball court.
“We used AI to scrape all of those (strong-mayor) rulings, and we developed a tool in order to classify it, and that’s one of the things that I think is important for people to think about when they think about AI,” Ellis said. “You can take a whole haystack of information, and will it give you the needle in the haystack? No, but it will give you a needle that’s covered in just a little bit of hay, and then the journalism can sort it out after that.”
The panel agreed that the journalism is in a precarious position and framed the news industry as “democratic infrastructure.” Christa Dickenson of the Canadian Public Affairs Channel, which recently went through staffing cuts, suggested the federal government should do more to incentivize a culture of philanthropy.
“In the months to come, we’ll be teasing out a little bit of a philanthropy strategy, and we hope to see Canadians get behind it with us,” Dickenson said of CPAC’s planning for future revenues.
‘The core value of journalism is boots on the ground, seeing things as it’s happening, and helping inform Canadians about the news, and empowering them to understand all of that,” Roy said. “Journalism, as a function, is important for a self-governing society, and it is part of our democratic infrastructure.’
— Nikita Roy, journalism & AI specialist
Hepfner questioned the critics of government funding for journalism.
“I don’t really understand this argument that journalism shouldn’t get government funding. Why not? The money has to come from somewhere. Is it better to have Jeff Bezos take over the Washington Post and get the money that way? You think he’s going to have less of a say over the editorial content than the government of Canada?”
Ellis suggested, from a personal perspective, that major players in the Canadian journalism industry prioritize innovation and share it with smaller players who can’t afford to experiment.
“We have to innovate, we have to have a better product than we do today, and we have to be realistic that we can’t . . . just continue to make a very similar product that we have in the past and still be relevant and still have the same sort of democratic function,” he said, suggesting that companies pool their resources to innovate AI tools.
“We don’t really compete with each other anymore . . . Go compete on a story, but don’t compete on what your technology platforms are, or some weird secrets that you think you have. Nobody cares. We have to be able to scale those tools,” Ellis said.
Despite AI being the main topic of the plenary discussion, all agreed that having reporters on the ground remains key to producing credible journalism.
“The core value of journalism is boots on the ground, seeing things as it’s happening, and helping inform Canadians about the news, and empowering them to understand all of that,” said Roy. “Journalism, as a function, is important for a self-governing society, and it is part of our democratic infrastructure.”
