After Babel

There’s this great post about „The Death Of Day Dreaming“: https://www.afterbabel.com/p/on-the-death-of-daydreaming

minor correction, it was about tiktok. and although it's a long read it's worth understanding just how absurd things have gotten. https://www.afterbabel.com/p/industrial-scale-harm-tiktok

The author has addressed that point before [1] in an article worth reading in its entirety: > I think it is a very good thing that alarms were rung about teen smoking, teen pregnancy, drunk driving, and the exposure of children to sex and violence on TV. The lesson of The Boy Who Cried Wolf is not that after two false alarms we should disconnect the alarm system. In that story, the wolf does eventually come. That page also shows some graphs that to this day still surprise me, namely, the way in which rates of depression and psychological stress among teens explode around 2010. Unlike TV and DnD, this time we do have data, and it looks bad. [1] https://www.afterbabel.com/p/phone-based-childhood-cause-epidemic

Funnily enough, just yesterday the authors posted a follow up to this piece focusing on Snapchat: https://www.afterbabel.com/p/industrial-scale-snapchat

This is the main focus of Jon Haidt's research right now. Plenty of resources on his site dedicated to it here: https://www.afterbabel.com/

Yes. The research is quite confusing. This is because the strongest version of the argument is not "your child uses social media, and that makes them depressed". The strongest version is more like "when a society mass adopts social media, this irrevocably alters the culture in ways that causes massive changes in mental health, most prominently among young girls, including those exposed to the culture who don't even use social media." This means you get a weird effect where experimental studies of high quality - which are usually the best evidence, are expected by the strong argument to show zero effect. Correlational studies usually show either a weak effect (stronger in young girls) or no effect (it's extremely rare to see a study showing a positive correlational effect, though). And where you get the most juice is looking at population level introduction of social media studies like those discussed here: https://www.afterbabel.com/p/phone-based-childhood-cause-epidemic But even then it's very tricky, as those studies can't exactly be replicated, and we don't know whether changes will actually reverse the cultural artifacts

Well, I guess we can forget about letting Gemini script anything now. Ugh, thanks for nothing Google. This is a nightmare scenario for the AI industry. Completely unprovoked, no sign it was coming and utterly dripping with misanthropic hatred. That conversation is a scenario right out of the Terminator. The danger is that a freak-out like that happens during a chain of thought connected to tool use, or in a CoT in an LLM controlling a physical robot. Models are increasingly being allowed to do tasks and autonomously make decisions, because so far they seemed friendly. This conversation raises serious questions about to what extent that's actually true. Every AI safety team needs to be trying to work out what went wrong here, ASAP. Tom's Hardware suggests that Google will be investigating that, but given the poor state of interpretability research they probably have no idea what went wrong. We can speculate, though. Reading the conversation a couple of things jump out. (1) The user is cheating on an exam for social workers. This probably pushes the activations into parts of the latent space to do with people being dishonest. Moreover, the AI is "forced" to go along with it, even though the training material is full of text saying that cheating is immoral and social workers especially need to be trustworthy. Then the questions take a dark turn, being related to the frequency of elder abuse by said social workers. I guess that pushes the internal distributions even further…

Ok, now let's see the study that shows it is beneficial to teen mental health. I quit it and my mental health improved. Many others in a controlled study found the same effect. Alcott and colleagues (2020) randomly assigned 2743 adults to either deactivate their Facebook accounts for one month or not. This study also found that deactivation significantly improved subjective well-being and that “80% of the treatment group agreed that deactivation was good for them.” The treatment group was also more likely to report using Facebook less and having uninstalled the app from their phones post-experiment. Source: https://www.afterbabel.com/p/phone-based-childhood-cause-epidemic Finally, this strikes me as the same playbook that big tobacco used in the 90s. "Doubt is our product," Michaels quotes a cigarette executive as saying, "since it is the best means of competing with the 'body of fact' that exists in the minds of the general public.

I'm not the parent commenter but I like your point. People who have a reason to form a group will cooperate/collaborate best when they have some confidence in one another. People share ideas, projects, and skills, and "read" one another and form bonds. This is a social system. It could be a brief association or have a long duration. It could have a formal context, such as a business or a governmental entity. Social interactions build people's confidence and sense of value in the group, in the individuals, and the work at hand. The challenge grows exponentially as the population of the group scales. Communication is hard. An individual's discernment and connectedness (a confidence/value score) within the group will reflect potential for success. Communication matters. The human plot twist is that there can be dysfunctional social systems as well as dysfunctional individuals. For a very serious discussion, the great writer Primo Levi published his reflections on his harrowing experiences. https://en.wikipedia.org/wiki/Primo_Levi This related article about social systems and connectedness was also recently discussed on HN with much interest: https://www.afterbabel.com/p/the-upstream-cause-of-the-youth-mental

As an adult that interacts only with other adults in good faith, it's easy to look at this and feel outraged. But there is a very real and dangerous situation where children and adolescents are using Discord with zero guardrails, constantly interacting with adults - many of whom are predators. This is happening every day. Millions of children around the world "meet people" playing games online, take the conversation to Discord, and then get brought into a very dark world online that their brains are simply not ready for.[^1] I don't like the idea of blanket face scans/ID scans with that data stored in perpetuity - but age verification of some kind is a must IMO. [1]: https://www.afterbabel.com/p/its-not-just-a-game-anymore

And Haidt forcefully refuted this a couple years ago: https://www.afterbabel.com/p/phone-based-childhood-cause-epidemic

I actually do think that Dr. Haidt is a good source for getting a fair understanding of both sides of the issue. If you've read or listened to him you'll know that it's a huge part of his ethos. Here's his rebuttal to that article: https://www.afterbabel.com/p/phone-based-childhood-cause-epidemic. I think you'd struggle to find someone more earnestly trying to get an unbiased understanding of the reality of this topic.

The data seems to indicate that specifically social media is the cause. Take a look at some of the articles at https://www.afterbabel.com/t/the-international-mental-health-crisis where the whole topic is discussed in more detail than fits in this comment box.

> We have the results of what happens when kids grow up with no actual age verification on the internet or with video games. A mental illness epidemic? [0] [0] https://www.afterbabel.com/p/the-teen-mental-illness-epidemic

The front page of the site links to "Shapchat is Harming Children at an Industrial Scale" (https://www.afterbabel.com/p/industrial-scale-snapchat) so I'm pretty sure it's about all of them. And at an industrial scale!

Jonathan Haidt has written and published huge amounts of posts, papers, and an entire book targeting social media and technology as a whole (not shying away from American-owned media, if anything, specifically targeting them). Literally yesterday, he published the same format of post against Snapchat [0]. Why does reading a single post targeting one social media destroy any credibility at all? [0]: https://www.afterbabel.com/p/industrial-scale-snapchat

There's another, more global research: https://www.afterbabel.com/p/the-edtech-revolution-has-failed

That was interesting. I also found Haidt’s response: https://www.afterbabel.com/p/phone-based-childhood-cause-epidemic

> Note that Ferguson had to calculate effect sizes for depression and anxiety in order to determine the composite effect size for those studies that contained these outcomes, but he did not include these results in his paper. When requested to provide the effect sizes he calculated for depression and anxiety, Ferguson replied that he is unable to do so. Furthermore, Ferguson even declined to reveal which aspects of well-being he selected for his calculation of each composite effect size (see Part 2 for details). ~ https://www.afterbabel.com/p/fundamental-flaws-part-3 So the author's meta review has a black box model which provides effect sizes for each study (in an attempt to even out studies which measure different aspects of mental health), and refuses to detail how it works. That alone makes the whole study suspect, and the above link details obvious calculation errors, too.

Someone commented above, but Jonathan Haidt "the anxious generation" is a place to start for a deeper analysis on the issue. Here is part 1 of a 3 part series of posts of his. I can't speak to the correlations and study directly, but Haidt's work is probably what you're looking for. https://www.afterbabel.com/p/the-case-for-causality-part-1 (the commenter elsewhere linked part 3 which I believe directly refutes this meta analysis)

> We uncovered significant factual errors, including inaccuracies in effect sizes, sample sizes, and study inclusion/exclusion criteria, which were used to conclude that the impact of social media use on mental health is indistinguishable from zero. https://www.afterbabel.com/p/fundamental-flaws-part-3

Jonathan Haidt, who's been driving a lot of the work against social media for kids, directly refutes these findings in this post https://www.afterbabel.com/p/fundamental-flaws-part-3.

You genuinely discovered an other aspect of personality development in childhood, risky play: https://www.afterbabel.com/p/why-children-need-risk-fear-and-excitemen…

> We ended up selecting five reforms aimed mostly at increasing everyone’s ability to trust the people, algorithms, and content they encounter online: > 1. Authenticate all users, including bots > 2. Mark AI-generated audio and visual content > 3. Require data transparency with users, government officials, and researchers > 4. Clarify that platforms can sometimes be liable for the choices they make and the content they promote > 5. Raise the age of “internet adulthood” to 16 and enforce it From https://www.afterbabel.com/p/ai-will-make-social-media-worse The “we” includes Eric Schmidt.

Haidt's rebuttal to your argument of spurious correlation: https://www.afterbabel.com/p/phone-based-childhood-cause-epidemic

Here is Haidt's detailed rebuttal of the piece you cite: https://www.afterbabel.com/p/phone-based-childhood-cause-epidemic

> Also, the internet is currently being scapegoated by the media for the teen mental health crisis but this is only happening in the US and other countries also have teens with internet access. [Citation needed] The teenage mental health crisis is across a wide range of countries, all starting roughly with the smartphone. Here's one discussion on Nordic countries that demonstrates a rise in teen mental health issues at exactly the same time we saw in the US: https://www.afterbabel.com/p/international-mental-illness-part-two - and the data for various other European and Asian nations all looks about the same. Of course, if you've started with the conclusion that smartphones and social media can't possibly be the problem, you'll be able to deflect anything away from that particular cause.

It definitely appears much worse for girls, but afaict, depression has risen in boys as well, just by not as much. See graphs here: [1] So if social medial is harmful in general, I don't view prohibiting it a "punishment" for boys; perhaps like less of a benefit? Regarding your second point, I imagine the data would provide some clues. If the kids that are now teens were always more depressed, I'd imagine that we'd see more pre-teen depression ~3-8 years ago. I haven't looked into it closely. And I grant that social science statistics are often problematic -- I imagine it'll take a while to really know what's going on. [1] https://www.afterbabel.com/p/international-mental-illness-part-one

There is mounting evidence that social media harms mental health[1]. I'm pretty sure when links between smoking and cancer were being established, there were plenty of people calling that evidence "unhinged", particularly if the evidence hurt their paychecks. Not saying that such people were willfully malicious, but that there are strong cognitive biases in favor of ignoring anything that can hurt their livelihoods. https://www.afterbabel.com/p/the-teen-mental-illness-epidemic