The digital landscape shaping the upcoming 2025 Canadian federal election feels unsettlingly unfamiliar, marked by an unprecedented surge of artificial intelligence-generated material filling a void left by traditional news sources on major platforms. Researchers describe the situation as bordering on the ‘dystopian,’ raising profound questions about information integrity as voters prepare to head to the polls on April 28th.
The Media Ecosystem Observatory (MEO), an Ontario-based group monitoring election-related online activity, reports a significant influx of AI-created content across social media. Aengus Bridgman, a key figure at MEO and a McGill University professor, identifies various forms, including simple memes and more sophisticated deepfake videos featuring politicians. Worryingly, the observatory also noted AI being used to impersonate legitimate news outlets, promoting fraudulent investment schemes, particularly those involving cryptocurrency. Even defunct community Facebook pages have been repurposed by AI for political advocacy, mimicking news dissemination where actual journalism has been blocked.
This phenomenon coincides directly with the aftermath of Bill C-18, the Online News Act, enacted in June 2023. The legislation mandated that tech giants like Meta compensate Canadian news outlets for content shared on their platforms. Meta’s response was swift and decisive: deeming the law ‘unworkable,’ it ceased news availability for Canadians on Facebook and Instagram in August 2023. This created what Bridgman terms an ‘online news void,’ a vacuum readily filled by AI-generated content, which he suggests Meta has failed to moderate effectively.
Meta, through spokesperson Julia Perreira, counters this narrative, stating the company invests heavily in election protection globally. Perreira highlighted a workforce of roughly 40,000 dedicated to safety and security, including 15,000 content moderators, backed by over $30 billion in technology and team investments over the past decade. Meta also pointed to its partnership with Elections Canada aimed at providing accurate voting information.
Despite the flood of synthetic content, Bridgman notes a silver lining: currently, there’s no concrete evidence suggesting Canadians are being successfully manipulated or swayed by falsehoods on a large scale. He credits voters’ growing awareness, stating, “Canadians are much more aware of and concerned about, for example, foreign interference, the role of deepfakes and manipulated content online than previously.” This awareness might explain MEO’s finding that Canadians are increasingly turning directly to the websites of established news organizations. Yet, the pervasiveness of AI deepfakes and impersonations begs the question: Do Canadians feel manipulated, even without definitive proof of widespread deception?
The MEO analysis, covering roughly 5,000 accounts across platforms like Facebook, Instagram, TikTok, YouTube, Bluesky, and Telegram, also revealed shifts in platform usage. Politicians and voters are exploring newer spaces like Bluesky and TikTok, with distinct party preferences emerging – Conservatives seeing slightly more engagement on X and Instagram, while Liberals dominate Bluesky. This fragmentation across platforms, Bridgman worries, could lead to divergent perceptions of key issues and potentially deepen political polarization.
The confluence of the online news void, the proliferation of difficult-to-verify AI content, and shifting platform dynamics creates a challenging environment for voters. While electors seem resilient so far, the sheer volume of synthetic material and its potential to mimic credible sources underscores a growing vulnerability. The situation implicitly highlights a critical need: should AI have a framework installed to validate sources, ensuring citizens can distinguish genuine information from sophisticated fabrications in an increasingly complex digital public square? Navigating this ‘dystopian’ feeling requires vigilance from voters and raises urgent questions about platform responsibility and the future of informed democratic participation.
References:
‘It feels dystopian:’ AI-generated content about federal election flooded online news void
