This audio is auto-generated. Please let us know if you have feedback.
The Australian government is moving ahead with a plan to make social media algorithms an opt-in option, which could give users more control over the content presented in their feeds.
In theory, this could reduce the amount of angst-inducing, rage-bait style content that confronts social media users every day. Yet Meta, and other social media providers, have repeatedly said that while users claim to prefer a non-algorithmic feed, engagement always declines with the alternative chronological display.
The question is, is that a bad thing?
This audio is auto-generated. Please let us know if you have feedback.
The Australian government is moving ahead with a plan to make social media algorithms an opt-in option, which could give users more control over the content presented in their feeds.
In theory, this could reduce the amount of angst-inducing, rage-bait style content that confronts social media users every day. Yet Meta, and other social media providers, have repeatedly said that while users claim to prefer a non-algorithmic feed, engagement always declines with the alternative chronological display.
The question is, is that a bad thing?
It’s certainly bad for social media platforms that make money from increased engagement in their apps. But from a social good perspective, it might be a net positive, in prompting people to spend less time scrolling through social media feeds.
First, in terms of impacts. There’s now an expansive body of evidencewhich indicates that algorithmic amplification can have harmful effects, including various academic reports that show that algorithms inherently drive more negative, divisive behaviors.
However, even if an alternative, algorithm-free experience means that divisive, angst-inducing content reaches fewer users, people might still be worse off due to content overwhelm and the echo chamber effect.
Algorithmic sorting was initially implemented in order to address issues with in-stream content overwhelm. In 2014, Facebook said it needed to implement algorithmic selection because people followed too many other people and Pages in the app.
As explained by Facebook: “[E]very time someone visits News Feed there are on average 1,500 potential stories from friends, people they follow and Pages for them to see, and most people don’t have enough time to see them all. These stories include everything from wedding photos posted by a best friend, to an acquaintance checking in to a restaurant.”
Given the breadth of each user’s expanded network, the algorithm enabled Facebook to ensure that people saw more of the posts they were most interested in, making certain that key updates from friends were not drowned out by brand posts and Page updates.
And it worked: “Previously, people read 57% of the stories in their News Feeds, on average,” Facebook said. “They did not scroll far enough to see the other 43%. When the unread stories were resurfaced, the fraction of stories read increased to 70%.”
Meta’s repeated tests over the years showed the same thing. When apps stop algorithmic ranking and instead show posts in chronological order, the number of stories people read and the likes and comments they submit decrease.
So algorithms drive a better user experience and keep people scrolling for longer. But they also amplify anger-inducing content, which could arguably be a key driver of the current division in political discourse. That has, of course, also created broader social divides.
Removing the algorithm could fix this. In addition, our collective understanding and usage of social media platforms is much more evolved now than it was in 2014, when people were less active in terms of curating and crafting their in-app experiences.
This has played out on Instagram and TikTok, where it’s now much harder to build an audience, because people are more discerning about the profiles they follow. Meanwhile, users are also much more active in unfollowing than they used to be.
But then again, that may also be due to advanced algorithmic sorting, which now shows users content based on their in-app activity. That means they don’t even need to follow anyone, because the system shows them more of what they like, based on their previous actions and indicators.
That kind of online ecosystem would also be impacted by an algorithm off switch, and it’s hard to predict how this might change user behaviors.
Either way, it seems like this could be worth an experiment, and a large-scale test in a smaller market like Australia could provide enough data to measure the impacts.
