A Hong Kong-owned website has created at least three fake AI journalists with Indigenous identities - including one posing as an Inuk reporter covering Nunavut - republishing rewritten articles from legitimate Canadian newsrooms without permission. The discovery, first reported by Nunatsiaq News, has drawn condemnation from media leaders who say the operation is siphoning both content and advertising revenue from real publishers on an industrial scale.
Leah Qammaniq is introduced to readers of We-News.com as an Inuk journalist raised in Nunavut, with a biography describing her writing as "warm, precise, and deeply local." She covers Inuit self-governance, food security, and daily life in the territory. She also does not exist. Qammaniq is one of 121 "AI correspondents" deployed by the site, which uses generated personas to repurpose journalism from outlets including CBC North, Nunavut News, and Nunatsiaq News.
CBC News subsequently identified two additional fabricated Indigenous journalists on the platform: Kaylee Johns, described as "a member of Yukon First Nation," and Warren Lafferty, described as "of Dene heritage."
The mechanics of industrial-scale content theft
We-News is owned by Hong Kong-based company AZ Medias Ltd., with a publisher listed as David Dragesco and a hosting provider in France. When Nunatsiaq News attempted to contact the site using its published email address, the message bounced. CBC News sent multiple emails and received no reply. Dragesco has been identified as a search-engine optimization expert whose separate business website, DiscoReady, is also published by AZ Medias Ltd.
The site's process, according to CBC News, consists largely of taking articles from legitimate news outlets, rewording them, and republishing the result under AI-generated bylines alongside AI-generated images. Paul Deegan, president of News Media Canada, which represents print and digital news publishers across the country, was direct in his assessment. "The theft is blatant and it's happening on an industrial scale," he told Nunatsiaq News.
Deegan said the only current remedy available to newsrooms is a lawsuit - an option that is neither affordable for most publishers nor straightforward when the accused party operates across multiple international borders. The financial damage extends beyond content theft. Corey Larocque, managing editor at Nunatsiaq News, told CBC News that the operation may be siphoning advertising revenue that would otherwise flow to legitimate newsrooms. "It's hard enough for news organizations to make a go of it without having an outside artificial intelligence news source using our own content to conceivably take away advertising dollars that could be ours," Larocque said.
How AI-generated headshots deceive readers
Researchers say the combination of AI-generated profile images and plausible biographical details gives invented journalists the appearance of legitimacy. Ahmed Al-Rawi, director of the Disinformation Project at Simon Fraser University, told Nunatsiaq News that human-looking profile images are the central mechanism of confusion. "The problem is that they actually have a picture of the alleged author of each news article, and that picture is a human, so it's confusing," Al-Rawi said. "One has to read the fine lines in order to figure it out."
We-News does include a disclosure stating it uses AI, but Al-Rawi told CBC News that a small-print transparency notice at the bottom of each article is not sufficient. "A lot of people might read the article … and think that it's real," he said. "The other problem is a lack of attribution, or at least clear attribution, to the original authors. It's very unfortunate that they depend on the original work of real journalists in order to generate revenue." The site's About Us page claims it "relies on artificial intelligence technologies" and that every story "is confirmed by multiple sources before publication."
Calls for stronger copyright protection
Brent Jolly, president of the Canadian Association of Journalists, described the challenge of combating these sites as "a game of whack-a-mole," noting that shutting down one operation does little when new ones can be created immediately, particularly when hosted in other jurisdictions. "As a journalist, it frankly disgusts me to see the public's right to know being manipulated by algorithms and artificial intelligence," he told Nunatsiaq News.
Jolly called on the federal government to strengthen copyright law to make the repurposing of stolen news content explicitly illegal, and said the public needs to become more media literate in the meantime. "This is about protecting the craft and upholding the public's right to know," he told CBC News. "We need to take this very seriously … Technology moves so fast and it's extremely difficult to keep up with, but we have to keep this in the driver's seat right beside us all the time."
Both Jolly and Deegan expressed disappointment that copyright protections for news content were absent from Canada's new AI strategy, released by the Liberal government earlier this summer. The strategy commits more than $2 billion in funding to expand the country's AI capacity, but does not address whether AI systems should be permitted to republish copyrighted journalism.
Why this matters for writers
This case illustrates a direct threat to the economic model that supports professional writing. When AI systems repurpose original journalism without attribution or compensation, they divert both audience attention and advertising revenue away from the newsrooms that produced the work. For writers, the fabricated personas add another layer of concern: AI-generated bylines with manufactured identities and biographies erode reader trust in authorship itself. Understanding how these systems operate - and the gaps in current copyright law that allow them to function - is becoming essential professional knowledge. Writers who want to navigate the intersection of AI and intellectual property can explore AI for Writers Courses or examine the underlying technology through Generative AI and LLM Courses that explain how large language models generate and repurpose text at scale.
Your membership also unlocks:
