Opinion: Michael Swanson.
I have, for a long time, had a real problem. I can’t help but comment on social media. I get into scraps, make sarcastic remarks, and generally get dragged into the absolute shit-show that is social media comments sections.
There was a time when disagreeing with someone online meant, at minimum, forming an argument. These days it’s often enough to slap a red angry-face reaction under a headline you haven’t read and move on with your afternoon. Late last year the Oxford English Dictionary named “rage bait” its word of the year, defined as content deliberately designed to provoke anger for the sake of engagement. That a dictionary felt the need to formalise the term tells you how mainstream the practice has become. In New Zealand’s small, tightly wound political ecosystem, it has reshaped what passes for public debate.
From debate to dopamine
Rage farming isn’t new, but it has matured into a business model. Platforms reward outrage because outrage keeps people scrolling, and pages which specialise in it are often chasing Facebook’s engagement payouts rather than any genuine political conviction. As one RNZ piece on the trend put it, a new wave of creators aren’t just inciting rage online, they’re profiting from it.
The incentive structure rewards the loudest, crudest take, not the most accurate one. A measured post explaining a policy trade-off gets a handful of likes. A post accusing a politician of being a “communist” or a “fascist” gets thousands of shares and a comment section that reads like a demolition site.
The Ardern years as a case study
Nobody in recent New Zealand political history absorbed more of this than Jacinda Ardern. Groups like Voices for Freedom, formed in late 2020, built large followings opposing lockdowns and vaccine mandates, and that audience became a durable base for broader anti-government sentiment that outlasted the pandemic itself. Billy Te Kahika’s Advance NZ page racked up more views than the Labour and National parties combined during the 2020 campaign before Facebook pulled it down for Covid misinformation.
The Disinformation Project, an independent research group, tracked the language used against Ardern daily from 2022 onward and found it becoming markedly more violent over time, homing in on terms like “horse”, “horseface” and threats invoking her young daughter.
This wasn’t robust criticism of the wage subsidy scheme or managed isolation. It was a sustained, gendered campaign of dehumanisation, amplified daily in comment sections and shared into Facebook groups that functioned as echo chambers rather than forums.
The right cops it too
It would be dishonest to pretend this only runs one direction. Centre-right figures get the same treatment from a different set of pages and commentariat, just with different vocabulary. David Seymour in particular has become a lightning rod, variously cast online as an architect of “monstrous” legislation or compared, not entirely seriously, to Donald Trump by critics during the Treaty Principles Bill debate.
Comment threads and Facebook groups opposing the coalition government regularly reach for words like “fascist” and worse, treating policy disagreements over intensification rules or charter schools as evidence of moral rot rather than differing views on governance. Luxon, too, gets rendered less as a Prime Minister with a tricky coalition to manage and more as a cardboard villain, “weak,” “feckless,” or simply mocked for his corporate background, depending on which group you scroll into.
The effect either way is the same: politicians become caricatures, and anyone who wants to discuss the actual substance of a bill gets shouted down as either a bootlicker or a traitor, depending on which side of the group they’ve wandered into.
Enter the machines
Generative AI has poured accelerant on all of this. RNZ reported on a page calling itself “NZ News Hub” that borrowed the name and branding of the defunct Newshub to push AI-generated images and video of real events, including a fabricated clip of the Prime Minister at Waitangi, racking up genuine engagement from followers who had no idea it wasn’t real.
AAP FactCheck has separately identified a network of Vietnam-linked disinformation pages pushing entirely invented stories, including an AI image of Luxon and Winston Peters supposedly silencing a crowd at the Vatican, and a fake claim that Luxon’s wife had personally wiped six-figure school lunch debt. Neither happened. Both looked convincing enough to fool people scrolling at speed.
Researchers at Victoria University have warned that New Zealand’s election rules simply aren’t ready for this, noting that AI-generated fake images of January’s Mt Maunganui landslide were shared widely during an actual national tragedy, and that even people who know an image is fake can still be swayed by it. That’s the genuinely alarming part. Debunking doesn’t fully undo the emotional residue a convincing fake leaves behind.
Where that leaves discourse
Pile all this together, rage-optimised pages, gendered pile-ons, cartoonish villainy applied to whichever politician a group dislikes this week, and a growing layer of AI-generated fabrication, and it’s little wonder that a 2023 Netsafe report found 91 percent of New Zealanders encounter misinformation at least monthly. The result isn’t robust debate, it’s parallel monologues. People aren’t arguing about the same set of facts anymore, because increasingly they aren’t even looking at the same facts.
To be fair, comment sections have always attracted the most excitable, most online sliver of the population, and it would be a mistake to treat them as a representative sample of what most New Zealanders actually think.
Most people are quietly getting on with their lives and don’t have an opinion on whether the Deputy Prime Minister is a “genuine, solid gold” name I won’t repeat here. But if the loudest, most visible layer of our political conversation is any guide to where public discourse is heading, then yeah, we’re completely buggered.
