How AI, Disinformation Distorts News and Politics

Canada's online sphere is warping politicians, whether they know it or not

How AI, Disinformation Distorts News and Politics
Find problems that you're the cause of. (Source: Daniil Komov via Unsplash)

Rory White, technology and democracy reporter for Canada's National Observer has recently released back-to-back stories revealing the dire straits of Canada's online ecosystem. The first, published on February 24, revealed that in a January presentation, Israeli company LogiVote promoted Victory, an AI-powered tool, that would scan for negative social media posts against a politician in order to create "digital fighters." As White reported, "Users are instructed to copy and paste an AI-generated comment onto social media, defending the candidate."

Boasting the tool's apparent ability to make a unique tailored response for each user, LogiVote founder Yair Chen told the National Observer "What we are doing is actually real democracy." This comes after Chen was quoted as promoting the Victory App fostering responses with "more efficiency and intelligence than any bot."

The NextCampaign Summit, where the LogiVote presentation took place, was co-founded by Harneet Singh. Singh was previously reported by the National Observer to be associated with an astroturfed community group in Caledon, ON. Despite this, prominent politicians and strategists like former Liberal House Leader Karina Gould, "Ontario's Minister of Transportation Prabmeet Sarkaria and conservative strategists Kory Teneycke and Hamish Marshall" attended the conference.

Those who study the political and AI space understandably rose alarms about this. One University of Concordia professor said the tool, presented as a boon for democracy, undermines "the very conditions that enable that democracy." Chen said that since the theoretical activist user would need to actively choose to participate, ethical issues were not present. Really convincing.

Deploying AI tools to fight online in the name of "democracy" is chilling enough, however White's following story solidified just how dangerous this environment is. According to the Media Ecosystem Observatory, just 100 users are responsible for nearly 70 per cent of all online conspiracy claims in Canada. Elon Musk's X was the largest congregation of these users. In an excellent dissection of how social media platforms prioritize engagement, White reports how politicians are responding to this perceived increase in conspiratorial thinking. Executive Director of Climate Caucus Zoe Grams warns that politicians, especially in smaller communities, are changing their language according to this conspiracy-laden environment.

Despite low engagement for most politicians on X, the prevailing attitude from government officials is that the platform remains a necessary tool for communication. But their continued use of X isn't neutral. As White notes in his story, Nature released a study that found X's algorithm shifted user opinion towards the right. As the study's authors note " the algorithm promotes conservative content and demotes posts by traditional media."

Our social media atmosphere is being weaponized specifically in this way. Outside of this latest report, Pivot previously found that political spam on TikTok was disproportionately pro-Conservative. As noted in the article, "0.2% of users in our sample are responsible for 15% of all comments, and nearly two-thirds of these accounts (65%) belong to the Conservative community." Under these circumstances, it's unsurprising that fake info attacking Mark Carney went viral in the last election.

Big Tech’s Siege on Journalism
Sustained attacks on news are a key part of the internet’s degradation

What does the right-wing disinformation consist of? It focuses on racist, xenophobic and anti-trans narratives. When companies like Meta ban Canadian news on their platform, disinformation like this rises to meet it. Memes and screenshots are now the primary avenue for information on Meta products. X is a hub for right-wing influencers and campaigns to explicitly move the Overton window to the right. Politicians, subsequently, are being fed with this ecosystem day-in and day-out; whether personally or through these ideas filtering into the feeds of their constituents.

It may be hard to track the direct effects of this hostile online environment, but the current attitude can still be gleaned by examining the current news and political situation. Conservative leader Pierre Poilievre phoned conservative partisan influencers after the last federal election to thank them for their work. The Conservative party created a new designation for content creator at their last convention. Weaponized anti-trans attacks flooded feeds after the Tumbler Ridge shooting and politicians largely ignored it. AI is supercharging disinformation around politics. It doesn't take a political genius to see where this is headed.

Yet there appears to be no political will to reign in AI or disinformation in at all. Carney's government pledged $1 billion to AI and quantum computing in the last budget. Minister of AI and Digital Innovation Evan Solomon was merely "disappointed" by OpenAI safety reps after it was revealed that relevant information about the Tumbler Ridge mass shooter's ChatGPT was kept from Canadian authorities. Solomon claimed "all options for us are on the table," but has made no meaningful moves to regulate AI before, even when Grok was revealed to be manufacturing non-consensual sexual deepfakes and child sexual abuse material (CSAM) en masse.

Meanwhile, AI apps like Victory are claiming to use AI models to create a "more efficient" army that politicians can deploy for their political campaigns. Rather than cut through the disinformation noise, explicitly partisan positions that support campaigns of candidates will only add more fuel to the fire.

Our politics are being affected directly by all this, but the implications for the news industry are likewise stark. Not only is there now a direct fight against AI disinformation in plain view, but the politicians and officials that outlets rely on for comment are being incubated in a right-wing propaganda engine. Ultimately, this propagandized environment will inevitably colour news coverage.

There's one simple way to describe the current situation: an absolute cluster fuck.