Aiden Lynch first started being exposed to toxic online content in high school.
But the now 21-year-old says he never bought into the more divisive ideas he saw there.
“I came from a background where my parents were always very involved, and they were always encouraging me to live according to my values,” said Lynch, who lives in Toronto.
“When I was exposed to this content, which I felt ran antithetical to my values, I found it … really divisive and disturbing,” he said, referring to content that espouses a hostile view toward women.
Now, as an adult, Lynch has become involved with White Ribbon, a Canadian non-profit focused on preventing gender-based violence. In September, White Ribbon launched a campaign that targets “algorithm sickness” — the gradual online exposure of boys and young men to misogynistic content.
White Ribbon’s campaign shifts the focus from a free speech perspective — whether misogynistic content should be allowed on social media platforms — to the responsibility of social media companies not to disseminate it.
“I think we used to focus on the content itself, and that’s why things got caught up in this free speech debate,” said Kaitlyn Regehr, a fellow at the University of Toronto’s Schwartz Reisman Institute for Technology and Society.
“I don’t think that’s the point. The point is whether Meta has the right to disseminate that content.”
The algorithmic pathway
Lynch says what alarmed him about the manosphere was how precisely it could target the insecurities some young boys and men have.
“It capitalizes on [gender] norms within society, which are quite dated, about what a man should be,” he said. “It’s really consumerist. It’s exceedingly unrealistic … like you should have this certain career, you should have this house, you should have those cars.”
He recalled how a friend in high school suggested he needed “two cars” to be attractive.
“Dude, you don’t need to have two cars. You don’t even drive yet,” Lynch remembers telling him. “If you talk to a girl in your class, she’s not going to ask you what car you’re driving. I’ve literally never been asked that in my life.”
The messages you hear online tend to be very superficial, Lynch adds. They do not encourage introspection, such as, “What can you do to make a relationship more solid?” he said.
Regehr notes that social media platforms’ algorithms are adept at recommending content that appears to provide viewers with an answer to their insecurities or frustrations.
“The digital space is very good at offering us solutions to feelings of loss of control,” said Regehr, who is also a researcher on digital literacy, youth and online radicalization at University College London, and the author of Smartphone Nation.
“Hate, harm and disinformation is often algorithmically prioritized because it’s more attention grabby than truth,” she added. “Harm or insult hook into our insecurities … and it’s that extra time [spent watching], that extra engagement that advertisers pay for.”
The concern, Regehr says, is not that one video suddenly radicalizes someone. Rather, it is that social media companies’ recommendation systems learn from each interaction and can gradually expose people to harmful ideas over time.
“You lingered on that post, and that gave your algorithm a seed to grow and feed you more of that content,” she said. “[Through] microdoses, content gets incrementally more extreme.”
Platform safeguards
Canadian Affairs conducted a time-limited test of TikTok’s system for recommending content.
It created a new account registered to a 16-year-old user in Canada. (TikTok did not request information on the user’s gender.)
It then conducted searches for “career advice for men,” “how to attract women” and “how to get fit for men.” It liked and bookmarked a dozen innocuous videos yielded from those searches.
For the first two days, the user’s feed showed a broad mix of self-improvement, fitness and dating content.
By the third day, the tone had shifted.
The account was shown increasingly adversarial and gendered dating content. This included videos advising men to respond to rejection with deceptive indifference, or to adopt a “villain” persona to attract women. One video implied that dating may sometimes involve minors, saying: “Sometimes you have to release your best catch [because] it isn’t legal.”
Some of the content crossed into explicitly misogynistic territory, portraying women as inherently untrustworthy and manipulative.
The feed was not uniformly adversarial. For example, a video by a female creator pushed back against “game” tactics and argued that only “toxic women” respond to them.
TikTok confirmed to Canadian Affairs that it uses machine learning, human safety moderators and other information to identify the likely age of its users. It does not rely solely on the age entered by users at sign-up.
TikTok also said it restricts content that may not be appropriate for teenagers, and interrupts recommendation patterns when otherwise permissible content — such as dieting and extreme fitness videos — becomes problematic through repeated exposure.
YouTube told Canadian Affairs its recommendation guardrails reduce the prominence of certain content for teen viewers, including videos involving hate and aggressive behaviour, dangerous acts and narrow body standards. A spokesperson also said the company works with child development experts to continually refine its systems.
Meta did not respond to multiple requests for comment. An April news release says the company’s Instagram “Teen Accounts” shield young users from age inappropriate content and creators.
Treating the ‘sickness’
White Ribbon’s “algorithm sickness” campaign treats exposure to online misogyny as something that can be interrupted before it escalates.
Lynch says the medical analogy is deliberate. “If we catch this early, then we can extirpate it,” he said.
That means starting conversations about misogyny before adolescence. Humberto Carolo, CEO of White Ribbon, says children need the tools to recognize manipulative and hateful ideas to “inoculate” them against such content.
For Lynch, intervention can also be as simple as challenging friends when they repeat harmful ideas.
“[I’d say] ‘Hey, dude, you can’t say that’,” he said. “It does ruin the vibe, right? But I feel like I have a responsibility to say something about it.”
Regehr says users can also influence what their algorithms learn by being deliberate about what they watch. This can include avoiding lingering on unwanted content and unfollowing accounts.
“We need to move away from being a passive consumer into being an active participant,” she said.
But she cautions against placing the burden entirely on individuals.
Canada’s proposed Safe Social Media Act, currently at second reading in the House of Commons, would require users to be at least 16 to create social media accounts. And it would require regulated services to assess and mitigate risks to children, among other things.
Companies that could demonstrate their platforms are safe for children could be exempt from the 16-year age minimum.
Regehr considers the shift in responsibility to platforms appropriate.
“If you’re unsafe, you have to exit the marketplace, and if you prove yourself to be safe again, you can re-enter the marketplace,” she said.
“That’s basically what we do with everything else — that’s what we do with food, it’s what we do with cars.”
