I think 70-80% of it is the business model, but the other 20-30% might just be baked into how it is.

Jonathan Haidt talks about how once social media usage became ubiquitous among teenagers around 2015 mental health problems began to skyrocket. And a big part of this was the algorithm serving up content designed to make people feel bad, but another part around feelings of being bullied turned out to largely be kids seeing their friends hanging out with each other without inviting them and this provoking feelings of alienation. That’s inevitable, I felt bad when I found out about parties or hang-outs I didn’t get invited to at that age as well. But I didn’t even know about 90% of them, and those I did I heard about through passing references rather than a stream of pictures and albums about how much fun everyone was having without me.

I think some level of a sense of isolation is inevitable under those circumstances, though I’m not sure that by itself would rise to the level of banning it outright. At least not before trying other interventions like addressing Meta’s “19 strikes before banning you for CSAM” rule. Kids are just the canaries in the coal mine here. Whatever these services are doing that is cooking developing brains is still turning up the heat on adult brains too, we can’t try to pretend we can be psychologically healthy engaging with something that we know is spiking depression and anxiety in our kids.

The culture of interacting just changed as more people got online and more tools became available to expand access to things. You used to just be able to have an unsecured comment section where anyone could come to your website and directly modify the page’s HTML and most of the time nothing would happen. You ought to have sanitized your inputs but there just wasn’t this background miasma that was going to flood your comment section full of spam, scans, and injecting malware into the page if you left an open text-entry box on the internet. Once it hit a certain scale and there was a certain amount of money in it then a lot of mess came with them.

Jonathan Haidt is someone who nobody should take seriously. Pretty much all of the data he cites is cherry-picked and the vast majority of people in trust and safety and similar will tell you that he is probably one of the least reliable authorities on this subject. He's aiming to sell fear, not to actually solve the problem.