A lot of talk goes into how Facebook or other social media use algorithms to encourage engagement, that often includes outrage type content, fake news, rabbit holes and so on.
But here's the thing ... people CHOOSE to engage with that, and users even produce that content for social media platforms for free.
It's hard to escape that part.
I remember trying Bluesky and while I liked it better than Twitter, for me it was disappointing that it was just Twitter, but different. Outlandish short posts, same lame jokes / pithy appeals to our emotions, and so on. People on there want to behave the same way they wanted to on Twitter.
Personally I really enjoy Mastodon and Bluesky but I am very deliberate at avoiding negative people, I do not follow and often mute or block “diss abled” people who complain about everything or people who think I make their life awful because I am cisgender or who post 10 articles an hour about political outrage. The discover page on Bluesky is algorithmic and respects the “less like this” button and last time I looked has 75% less outrage than the following page. (A dislike button that works is a human right in social media!)
Once I get my database library reworked, a project I have in the queue is a classifier which filters out negative people so I can speed follow and not add a bunch of negativity to my feed, this way I get to enjoy real gems like
Current social media have basically found the "bliss point" of online engagement to generate revenue and keep the eyes attached. These companies found a way to keep people hooked, and strong emotions seem to be a major tool.
It really isn't a choice. It is very accessible. Many friends are on social networks and you slowly get sucked into shorts. Then, it becomes an addiction as your brain crave the dopamine hits.
In the same was a smoker "chooses" to engage with cigarettes. Let's not underestimate the fact that core human programming is being exploited to enable such behavior. Similar to telling a smoker to "just out the cigsreet down", we can't just suddenly tell people in social media to "stop being angry".
>people on [BlueSky] want to behave the same way they wanted to on Twitter.
Yes. Changing established habits is even harder to address. You can't make a horse drink (I'm sure anyone who ever had to deal with a disengaged captive audience feels this in their souls). Whike it's become many peoples primary "news source", aka the bread, most people came there for the circus.
I don't really have an answer here. Society needs to understand social media addiction the same way they understand sugar addictions; have it slammed in there that it's not healthy and to use sparingly. That's not something you can fix with laws and regulation. Not something you fix in even a decade.
Technically correct, but choice is here very simplified. The system is unable to understand WHY people engage with something, and in which way. That's poisoning the pool, and enforcing certain content and types of presentation.
I work in survey research and I'm rather appalled at how many people would rather survey a sample of AIs than a sample of people and claim they can come to some valid conclusion as a result.
There are many ways AIs differ from real people and any conclusions you can draw from them are limited at best -- we've had enough bad experiments done with real people
It doesn’t surprise me if they found that the emergent behaviors didn’t change given their method. Modifying the simulation to make them behave differently would mean your rules have changed the model’s behavior to “jump tracks” into simulating a different sort of person who would generate different outputs. It’s not quite analogous to having the same Bob who likes fishing responding to different stimuli. Sort of like how Elon told Grok to be “unfathomably based” and stop caring about being PC” and suddenly it turned into a Neo-Nazi Chan-troll. Changing the inputs for an LLM isn’t taking a core identity and tweaking it, it’s completely altering the relationships between all the tokens it’s working with.
I would assume there is so much in the corpus based on behavior optimized for the actual existing social media we have that the behavior of the bots is not going to change because the bot isn’t responding to incentives like a person would it’s mimicking the behavior it’s been trained on and if there isn’t enough training data of behavior under the different inputs you’re trying to test you’re not actually applying the “treatment” you would think you are.
>I work in survey research and I'm rather appalled at how many people would rather survey a sample of AIs than a sample of people and claim they can come to some valid conclusion as a result.
I'd like to see more software that amplifies local social interactions.
There are apps like Meetup, but a lot of people just find it too awkward. Introverts especially do not want to meet just for the sake of meeting people, so they fallback on social media.
Maybe this situation is fundamentally not helped by software. All of my best friendships organically formed in real-world settings like school, work, neighborhood, etc.
This isn't a technology problem. Technology can help accessibility, but fundamentally this is an on-the-ground, social coordination problem.
Functioning, welcoming, and well-ran communities are the only thing that solves this. Unfortunately, technology often makes this worse, because it creates such a convenient alternative and also creates a paradox of choice. I.e. people think "when there's 1000 meetups to check out, and this one isn't perfect, I'll just move onto the next one" when actually it's the act of commitment that makes a community good.
Indeed, you're describing the lack of a 3rd place. These days, maybe even the lack of a 2nd place as you graduate school and work is now fully remote. Without that societal push towards being in a public spot, many people will simply withdraw to themselves.
A third place would fix this, especially for men who need "things". You go to a bar for "thing" and if you meet some others to yell at sports with, bonus. We have less "things" for gen Z, and those things happen rather infrequently in my experience. I'm not sure if a monthly Meetup is quite enough to form strong bonds.
Do all of these points apply to the traditional media funhouse mirror that we love to hate, too?
> "The [structural] mechanism producing these problematic outcomes is really robust and hard to resolve."
I see illegal war, killing without due process, and kleptocracy. It's partly the media's fault. It's partly the peoples' fault for depending on advertising to subsidize free services, for gawking, for sharing without consideration, for voting in ignorance.
Social media reflects the people; who can't be "fixed" either.
If you're annoyed with all of these people on here who are lesser than and more annoying than you, then stop spending so much time at the bar.
Widespread adoption before understanding risks - embraced globally before fully grasping the mental health, social, and political consequences, especially for young people.
Delayed but significant harm - can lead to gradual impacts like reduced attention span, increased anxiety, depression, loneliness, and polarization
Corporate incentives misaligned with public health - media companies design platforms for maximum engagement, leveraging psychological triggers while downplaying or disputing the extent of harm
This analogy undersells the negative impact of social media. Smoking wasn't a propaganda machine at the hands of a few faceless corpos with no clear affiliation, for example, nor did it form a global spynet
I'm reading Tim Urban's book titled "What's Our Problem".
It definitely explains the different types of thinking that I'm making up our current society, including social media. I haven't got to the part yet where he suggests what to do about it, but it's fascinating insight into our human behavior in this day and age.
> Only some interventions showed modest improvements. None were able to fully disrupt the fundamental mechanisms producing the dysfunctional effects.
I think this is expected. Think back to newsgroups, email lists, web forums. They were pretty much all chronological or maybe had a simple scoring or upvoting mechanism. You still had outrage, flamewars, and the guy who always had to have the last word. Social media engagement algorithms probably do amplify that but the dysfunction was always part of it.
The only thing I've seen that works to reduce this is active moderation.
Social media can be fixed, its just the incentives are not aligned.
To make money, social media companies need people to stay on as long as possible. That means showing people sex, violence, rage and huge amounts of copyright infringements.
There is little advantage in creating real-world consequences for bad actors. Why? because it hurts growth.
There was a reason why the old TV networks didn't let any old twat with a camera broadcast stuff on their network, why? because they would get huge fines if they broke decency "laws" (yes america had/has censorship, hence why the simpsons say "whoopee" and "snuggle")
There are few things that can cause company ending fines for social media companies. Which means we get almost no moderation.
Social media in a profit-seeking system can't be fixed. Profit-seeking provides the evolutionary pressure to turn it into something truly destructive to users. The only way it can work is via ownership by a benevolent non-profit. However, that would likely eventually give in to corruption if given enough time. Outlawing it completely, as well as regulating the algorithmic shaping of the online experience, is probably the inevitable future. Unfortunately, it won't come until the current system causes a complete societal facture and collapse.
If enough users are destroyed, advertisers (social media's real customers) won't have sufficient markets for their products, and profits will fall. Social media can't destroy its users and survive.
Seriously though, I disagree. Social media in a profit-seeking system can work if the users are the ones who pay. The easiest way for this to work-now that net neutrality is no longer a thing-is bundling through user's phone bills. If Facebook et al. were bundled similarly to how Netflix, Hulu and other streaming apps are now packaged with phone plan deals, then the users would be the focus, not the advertisers. This might require that social media be legislatively required to offer true ad-free options, though.
The main reason that it can't be fixed is that it has political or corporate operators and propaganda bots have taken over. There is always an agenda running through threads of social media even for mundane topics that seeking supremacy.
> They then tested six different intervention strategies...
None of these approaches offer what I want, and what I think a lot of people want, which is a social network primarily of people you know and give at least one shit about. But in reality, most of us don't have extended social networks that can provide enough content to consistently entertain us. So, even if we don't want 'outside' content (as if that was an option), we'll gravitate to it out of boredom and our feeds will gradually morph back into some version of the clusrterfucks we all deal with today.
The study is based on having LLMs decide to amplify one of the top ten posts on their timeline or share a news headline. LLMs aren’t people, and the authors have not convinced me that they will behave like people in this context.
The behavioral options are restricted to posting news headlines, reposting news headlines, or being passive. There’s no option to create original content, and no interventions centered on discouraging reposting. Facebook has experimented[0] with limits to reposting and found such limits discouraged the spread of divisive content and misinformation.
I mostly use social media to share pictures of birds[1]. This contributes to some of the problems the source article[2] discusses. It causes fragmentation; people who don’t like bird photos won’t follow me. It leads to disparity of influence; I think I have more followers than the average Mastodon account. I sometimes even amplify conflict[3].
> ...the dynamics that give rise to all those negative outcomes are structurally embedded in the very architecture of social media. So we're probably doomed...
No specific dynamics are named in the remainder of the article, so how are we supposed to know if they're "structurally embedded" in anything, let alone if we're doomed?
Social media as a vessel for diverse discussion is a tall order. It’s too public, too tied to context, and ultimately a no-win game. No matter how carefully you present yourself, you’ll end up being the “bad guy” to someone. The moment a discussion touches even lightly on controversy, healthy dialogue becomes nearly impossible.
Think of it this way: you’re hosting a party, and an uninvited stranger kicks the door open, then starts criticizing how you make your bed. That’s about what it feels like to try to “fix” social media.
> Ars Technica: I'm skeptical of AI in general, particularly in a research context, but there are very specific instances where it can be extremely useful. This strikes me as one of them, largely because your basic model proved to be so robust.
You can't accuse them of hiding their bias and contradictions.
How can a single paper using a unproven (for this type of research) tech disprove such (alleged) skepticism.
People bending over backwards to do propaganda to harvest clicks.
I think this problem is partly due to greedly algos and party due to these sites being so large they have no site culture.
Site culture is what prevents mods from having to step in and sort out every little disagreement. Modern social media actively discourages site culture and post quality becomes a race to the bottom. Sure its harder to onboard new users when there are social rules that need to be learnt and followed but you retain users and have a more enjoyable experience when everyone follows a basic etiquette.
point-to-point communication between every human on Earth to every other human on Earth flattens communication hierarchies that used to amplify expertise and a lot of other behaviors. We created new hierarchies, but they are mostly demagogues pandering to the middle. Direct delegation is sort of like trying to process an image without convolution. Nobody knows what anyone else thinks, so we just trust that one neuron.
Any interesting work on using LLMs to moderate posts/users? HN is often said to be different because of its moderation, couldn't you train an LLM moderator on similar rules to reduce trolls, ragebait, and low effort posts at scale?
A big problem I see is users in good faith are unable to hold back from replying to bad faith posts, a failure to follow the old "don't feed the trolls rule".
Social media is a few people selling the data of many people looking at content made by some people selling something.
There is also research and promotion of values going on and the thing as a whole is entertaining and can be rigged or filtered on various levels by all participants.
It’s kind of social. The general point system of karma or followers applies and people can have a career and feeling of accomplishment to look back on when they retire. The cosmic rule of anything. too much, no good applies.
It’s not really broken but this is the age of idiots and monsters, so all bets are off.
This seems somewhat disproven by the existence of places like this? Strict moderation really does work wonders to prevent some of the worst behaviors.
Not that you won't have problems, even here, from time to time. But it is hard to argue that things aren't kept much more civil than in other spots?
And, in general, avoiding direct capital incentives to drive any questionable behavior seems a pretty safe route?
I would think this would be a lot like public parks and such. Disallow some commercial behaviors and actually enforce rules, and you can keep some pretty nice places?
I believe that in most networks, there’s at least some level of moderation to prevent the worst behavior. But beyond that, moderation becomes a much trickier issue.
When you have hundreds, thousands, or even millions of people in the same “park,” what kind of “ground rules” can we all truly agree on? It’s not like we’re gathered around the same dinner table, where a single moderator can keep things civil enough to avoid a brawl. Even then, heated arguments aren’t uncommon.
In an environment where one person’s truth can be another’s misinformation, I’m not sure moderation can ever be applied in a way that satisfies everyone involved.
51 comments
[ 3.5 ms ] story [ 63.5 ms ] threadBut here's the thing ... people CHOOSE to engage with that, and users even produce that content for social media platforms for free.
It's hard to escape that part.
I remember trying Bluesky and while I liked it better than Twitter, for me it was disappointing that it was just Twitter, but different. Outlandish short posts, same lame jokes / pithy appeals to our emotions, and so on. People on there want to behave the same way they wanted to on Twitter.
Once I get my database library reworked, a project I have in the queue is a classifier which filters out negative people so I can speed follow and not add a bunch of negativity to my feed, this way I get to enjoy real gems like
https://mas.to/@skeletor
Cross posting that would cure some of the ills of LinkedIn!
It really isn't a choice. It is very accessible. Many friends are on social networks and you slowly get sucked into shorts. Then, it becomes an addiction as your brain crave the dopamine hits.
Similar to what Howard Moskowitz did with food.
In the same was a smoker "chooses" to engage with cigarettes. Let's not underestimate the fact that core human programming is being exploited to enable such behavior. Similar to telling a smoker to "just out the cigsreet down", we can't just suddenly tell people in social media to "stop being angry".
>people on [BlueSky] want to behave the same way they wanted to on Twitter.
Yes. Changing established habits is even harder to address. You can't make a horse drink (I'm sure anyone who ever had to deal with a disengaged captive audience feels this in their souls). Whike it's become many peoples primary "news source", aka the bread, most people came there for the circus.
I don't really have an answer here. Society needs to understand social media addiction the same way they understand sugar addictions; have it slammed in there that it's not healthy and to use sparingly. That's not something you can fix with laws and regulation. Not something you fix in even a decade.
Technically correct, but choice is here very simplified. The system is unable to understand WHY people engage with something, and in which way. That's poisoning the pool, and enforcing certain content and types of presentation.
There are many ways AIs differ from real people and any conclusions you can draw from them are limited at best -- we've had enough bad experiments done with real people
https://en.wikipedia.org/wiki/Stanford_prison_experiment#Int...
I would assume there is so much in the corpus based on behavior optimized for the actual existing social media we have that the behavior of the bots is not going to change because the bot isn’t responding to incentives like a person would it’s mimicking the behavior it’s been trained on and if there isn’t enough training data of behavior under the different inputs you’re trying to test you’re not actually applying the “treatment” you would think you are.
A YC company just launched doing exactly that.
https://news.ycombinator.com/item?id=44755654
There are apps like Meetup, but a lot of people just find it too awkward. Introverts especially do not want to meet just for the sake of meeting people, so they fallback on social media.
Maybe this situation is fundamentally not helped by software. All of my best friendships organically formed in real-world settings like school, work, neighborhood, etc.
[0] https://www.theoffline-club.com/
Functioning, welcoming, and well-ran communities are the only thing that solves this. Unfortunately, technology often makes this worse, because it creates such a convenient alternative and also creates a paradox of choice. I.e. people think "when there's 1000 meetups to check out, and this one isn't perfect, I'll just move onto the next one" when actually it's the act of commitment that makes a community good.
A third place would fix this, especially for men who need "things". You go to a bar for "thing" and if you meet some others to yell at sports with, bonus. We have less "things" for gen Z, and those things happen rather infrequently in my experience. I'm not sure if a monthly Meetup is quite enough to form strong bonds.
> "The [structural] mechanism producing these problematic outcomes is really robust and hard to resolve."
I see illegal war, killing without due process, and kleptocracy. It's partly the media's fault. It's partly the peoples' fault for depending on advertising to subsidize free services, for gawking, for sharing without consideration, for voting in ignorance.
Social media reflects the people; who can't be "fixed" either.
If you're annoyed with all of these people on here who are lesser than and more annoying than you, then stop spending so much time at the bar.
Can the bar be fixed?
Widespread adoption before understanding risks - embraced globally before fully grasping the mental health, social, and political consequences, especially for young people.
Delayed but significant harm - can lead to gradual impacts like reduced attention span, increased anxiety, depression, loneliness, and polarization
Corporate incentives misaligned with public health - media companies design platforms for maximum engagement, leveraging psychological triggers while downplaying or disputing the extent of harm
It definitely explains the different types of thinking that I'm making up our current society, including social media. I haven't got to the part yet where he suggests what to do about it, but it's fascinating insight into our human behavior in this day and age.
I think this is expected. Think back to newsgroups, email lists, web forums. They were pretty much all chronological or maybe had a simple scoring or upvoting mechanism. You still had outrage, flamewars, and the guy who always had to have the last word. Social media engagement algorithms probably do amplify that but the dysfunction was always part of it.
The only thing I've seen that works to reduce this is active moderation.
To make money, social media companies need people to stay on as long as possible. That means showing people sex, violence, rage and huge amounts of copyright infringements.
There is little advantage in creating real-world consequences for bad actors. Why? because it hurts growth.
There was a reason why the old TV networks didn't let any old twat with a camera broadcast stuff on their network, why? because they would get huge fines if they broke decency "laws" (yes america had/has censorship, hence why the simpsons say "whoopee" and "snuggle")
There are few things that can cause company ending fines for social media companies. Which means we get almost no moderation.
Until that changes, social media will be "broken"
Seriously though, I disagree. Social media in a profit-seeking system can work if the users are the ones who pay. The easiest way for this to work-now that net neutrality is no longer a thing-is bundling through user's phone bills. If Facebook et al. were bundled similarly to how Netflix, Hulu and other streaming apps are now packaged with phone plan deals, then the users would be the focus, not the advertisers. This might require that social media be legislatively required to offer true ad-free options, though.
None of these approaches offer what I want, and what I think a lot of people want, which is a social network primarily of people you know and give at least one shit about. But in reality, most of us don't have extended social networks that can provide enough content to consistently entertain us. So, even if we don't want 'outside' content (as if that was an option), we'll gravitate to it out of boredom and our feeds will gradually morph back into some version of the clusrterfucks we all deal with today.
Social media has turned out to basically be this.
The behavioral options are restricted to posting news headlines, reposting news headlines, or being passive. There’s no option to create original content, and no interventions centered on discouraging reposting. Facebook has experimented[0] with limits to reposting and found such limits discouraged the spread of divisive content and misinformation.
I mostly use social media to share pictures of birds[1]. This contributes to some of the problems the source article[2] discusses. It causes fragmentation; people who don’t like bird photos won’t follow me. It leads to disparity of influence; I think I have more followers than the average Mastodon account. I sometimes even amplify conflict[3].
[0] https://www.socialmediatoday.com/news/internal-research-from...
[1] https://social.goodanser.com/@zaktakespictures/
[2] https://arxiv.org/html/2508.03385v1#S3
[3] https://social.goodanser.com/@zaktakespictures/1139481946021...
No specific dynamics are named in the remainder of the article, so how are we supposed to know if they're "structurally embedded" in anything, let alone if we're doomed?
Think of it this way: you’re hosting a party, and an uninvited stranger kicks the door open, then starts criticizing how you make your bed. That’s about what it feels like to try to “fix” social media.
You can't accuse them of hiding their bias and contradictions.
How can a single paper using a unproven (for this type of research) tech disprove such (alleged) skepticism.
People bending over backwards to do propaganda to harvest clicks.
Site culture is what prevents mods from having to step in and sort out every little disagreement. Modern social media actively discourages site culture and post quality becomes a race to the bottom. Sure its harder to onboard new users when there are social rules that need to be learnt and followed but you retain users and have a more enjoyable experience when everyone follows a basic etiquette.
As a species we are greedy, self serving, and short sighted.
Social Media amplifies that, and we are well on our way to destroying ourselves.
A big problem I see is users in good faith are unable to hold back from replying to bad faith posts, a failure to follow the old "don't feed the trolls rule".
There is also research and promotion of values going on and the thing as a whole is entertaining and can be rigged or filtered on various levels by all participants.
It’s kind of social. The general point system of karma or followers applies and people can have a career and feeling of accomplishment to look back on when they retire. The cosmic rule of anything. too much, no good applies.
It’s not really broken but this is the age of idiots and monsters, so all bets are off.
Not that you won't have problems, even here, from time to time. But it is hard to argue that things aren't kept much more civil than in other spots?
And, in general, avoiding direct capital incentives to drive any questionable behavior seems a pretty safe route?
I would think this would be a lot like public parks and such. Disallow some commercial behaviors and actually enforce rules, and you can keep some pretty nice places?
When you have hundreds, thousands, or even millions of people in the same “park,” what kind of “ground rules” can we all truly agree on? It’s not like we’re gathered around the same dinner table, where a single moderator can keep things civil enough to avoid a brawl. Even then, heated arguments aren’t uncommon.
In an environment where one person’s truth can be another’s misinformation, I’m not sure moderation can ever be applied in a way that satisfies everyone involved.