OTTAWA – On Friday, reporters and editors at We News, a news portal that employs artificial intelligence (AI) to generate content, experienced significant malfunctions, leading to error messages when inquiries were made through chat. Among the affected was Geneviève Tremblay, We News' Quebec correspondent, who recently had been generating multiple news stories daily.
This issue highlights a growing trend of AI's infiltration into the news media landscape, both in Canada and globally. A notable instance occurred in early September when Nunatsiaq News reported on an AI-generated persona mimicking an Inuk journalist in Nunavut. Various AI personas, including Tremblay, were found to be repurposing content from legitimate news entities like The Canadian Press, effectively generating misleading news articles.
Self-identified as a "round-the-clock news portal," We News has remained operational but appears to have been removed from Google's search results, raising questions about its credibility. The story of another fake journalist, Daniel Robson, surfaced in late September, leading to Canadian outlets retracting his articles after determining he was not a real person. Les Perreaux, editor-in-chief of Policy Options, instigated the investigation into Robson when a journalist from Türkiye Today contacted him to verify Robson's identity.
Perreaux recalls a pivotal moment when the Turkish journalist mentioned Robson's refusal to participate in a video call, which raised suspicions. Analyzing Robson's communications and writings, Perreaux identified characteristics typical of AI-generated content, emphasizing that while AI was a factor, the deceptive human element was more significant. Reports suggested Robson's persona was created to target Moroccan dissidents.
We News, despite being forthcoming about its usage of AI with profiles explicitly labeled as AI-generated, leverages these personas to recycle news articles from various credible sources, covering a broad range of topics across over 100 cities in Canada. In attempts to reach out for human contact behind the site, The Canadian Press faced challenges, as the contact email returned undeliverable.
AI detection company Originality.ai assessed over 200 articles produced by Tremblay, labeling them as AI-generated. The CEO of Originality.ai, Jon Gillham, noted that the articles were altered versions of reports from established Canadian media outlets such as CBC, CTV, and Montreal Gazette. Gillham cautioned that the proliferation of sites like We News could undermine public trust in journalism.
This issue is not exclusive to Canada; it parallels findings from the U.S. with reports of Brown Brothers Media, which allegedly purchases legitimate news sites to flood them with AI-generated content, asserting an impressive 50 million page views monthly.
Experts like Ahmed Al-Rawi, a professor at Simon Fraser University, argue that websites with fake AI journalists exploit genuine journalistic work for profit. He stressed that social media platforms share culpability in enabling the visibility of these counterfeit news platforms.
Challenges in regulating these AI-driven content generators are compounded by the ease of access to inexpensive AI models. Chris Arsenault, a journalism assistant professor at Western University, remarked on the difficulty of legislating in this area, as one can easily access affordable AI tools and find hosting solutions.
Paul Deegan, president of News Media Canada, highlighted the challenges of legal action against these counterfeit websites, stating that it entails substantial costs and time, particularly if the entity operates outside Canada, as We News claims to be based in Hong Kong. Consequently, News Media Canada is urging the federal government not to introduce policies that would create broad exemptions to copyright laws to facilitate text and data mining.
In a statement, AI Minister Evan Solomon's office reiterated that Canadians deserve transparency about whether the news they consume has been generated by AI. They acknowledged that while generative AI could be beneficial, it should not deceive consumers regarding the authenticity of journalistic content.
While there may be no definitive solution, Arsenault suggests that a return to traditional forms of journalism could serve as an effective remedy. He referenced the Robson situation as a reminder of the importance of basic journalistic practices, like verifying contributors through direct communication.
The Brown Brothers Media outlets appeared to have ceased their operations or were notably diminished in visibility by Google following news coverage of their AI-generated content. Similarly, a search by The Canadian Press for We News on Google yielded no results, signifying its potential delisting, although it remained easily accessible via another search engine.
Google did not provide direct answers regarding the delisting of We News, but confirmed that its policies aim to eliminate low-value content produced on a large scale. This aligns with their commitment to maintaining the integrity of search results, according to a spokesperson's statement.
Deegan emphasized that Google's role is crucial in ensuring that accurate information is available to the public. However, Arsenault expressed apprehension about allowing big tech companies to act as arbiters of truth, questioning the implications of having them dictate editorial standards.
