Posts / social-media
Turning Off the Algorithm: A Nice Idea That Might Not Fix Anything
Canberra’s floated another one this week: opt-out laws for social media algorithms, letting people switch off the recommendation engine and go back to a plain chronological feed. Platforms would also be on the hook for “minimising and preventing harm.” It’s the kind of headline that makes you nod along at first, then start picking at the threads once you actually think about how these things work in practice.
I’ve spent a fair chunk of my career around systems that decide what to show people and in what order, so I don’t come at this cold. And my honest first reaction was relief, followed pretty quickly by scepticism. Relief because someone in government has finally noticed that the “for you” page is not a neutral mirror of your interests, it’s a machine tuned to keep you scrolling. Scepticism because turning a dial from “on” to “off” assumes the dial is that simple. It isn’t.
The discussion I stumbled into online this week made the point better than I could. One bloke described an account he’s used for a decade, nothing but yoga and secondhand baby clothes, and somehow his feed is wall-to-wall Pauline Hanson and geopolitics. Someone else pointed out you can already switch Facebook and Instagram to a chronological, friends-only view, it’s just buried three menus deep because nobody at Meta wants you using it. The reality, as one commenter put it, is that your mates don’t post enough content to keep you dopamine-drip-fed for three hours a day, so there’s zero commercial incentive to make that the default. That’s the bit that gets me. The technology to give people a calmer, less manipulative experience already exists. It’s just not profitable, so it’s hidden.
Then there’s the YouTube problem, which came up a lot in the same thread and which I think about most days. A few people swore by watching everything logged out, searching for exactly what they want, closing the tab, no algorithm, no rabbit hole. Fair enough, and I half do that myself. But someone else made the point that a brand-new YouTube account with no algorithm at all is close to useless, just noise, no way to find anything good. So the honest answer is probably “it depends what you’re using it for,” which is not a satisfying answer but is the true one. I don’t think there’s a single lever here. Recommendation and curation aren’t inherently evil, they’re just currently optimised for engagement instead of for your wellbeing, and those are different things that happen to look similar for about the first ten minutes of scrolling.
What actually annoyed me about the thread, in a constructive way, was the bit about “not interested” and “don’t recommend this channel” being quietly ignored a week later. That’s not a design flaw, that’s a decision. Someone at these companies looked at the data showing users don’t want certain content and decided to keep feeding it anyway because it still generates engagement, even negative engagement. Rage is sticky. My daughter, sixteen and no fool, has told me more than once that she’s flagged the same rubbish on her feed a dozen times and it keeps crawling back like a cane toad. She’s not naive about any of this, her generation grew up fluent in exactly how they’re being played, and yet she still opens the app every day because that’s where her friends are. That’s the trap in a sentence: understanding the mechanism doesn’t switch off the pull of it.
So will an opt-out law actually change much? Genuinely, I don’t know. My guess is that most people won’t opt out, the same way most people never touch the privacy settings buried in their phone, and the platforms know that too. Defaults are destiny. If the law forces a genuinely equal, unhidden, one-tap choice, and holds companies liable for harm minimisation with actual teeth rather than a “please explain” letter once a year, it could matter. If it becomes another checkbox nobody clicks while the engagement engine keeps humming underneath, it’s theatre. I’d like to be wrong about that.
Where I land, and I’ll allow myself one sentence without a joke attached to soften it: it’s genuinely good that regulators are treating attention as something worth protecting, even if this particular bill turns out to be a blunt instrument. We’ve spent fifteen years letting a handful of companies run the biggest unregulated psychology experiment in human history, and “let people flip a switch” is a modest start, not a solution. I’ll take modest starts over none. Just don’t expect the toggle to save you from yourself.