Yudkowsky said it is known observationally that something about social media is harming people quite a lot, children more than adults.

Judged as of 16 March 2024 against the observational research on social media use and wellbeing. TRUE: authoritative reviews and bodies (e.g. the 2023 US Surgeon General advisory, major systematic reviews) treat observational evidence of substantial harm from social media as established, with consistently larger associations for children and adolescents than for adults. MOSTLY TRUE: observational studies consistently show associations between social media use and worse mental health, stronger in young people, but the effects are modest rather than 'quite a lot', or causality remains disputed while the evidence is still treated as a real concern. MIXED: one part holds and the other does not. For example, associations are documented, but there is no consistent evidence they are stronger in children than in adults, or the size of the harm is sharply disputed among leading researchers. MOSTLY FALSE: the observational associations found are small or inconsistent (e.g. Orben & Przybylski-style findings of negligible effects dominate), so 'known to be harming people quite a lot' overstates a weak correlation. FALSE: the observational evidence shows no association, or one that is stronger in adults than in children.

Set before the research began.

1. E1 Adolescent Mental Health in the Digital Age: Facts, Fears and Future Directions PubMed Central (author manuscript of Odgers & Jensen, J Child Psychol Psychiatry)

   “To summarize, a review of meta-analytic work, large-scale preregistered studies, and intensive daily and momentary assessments provides little evidence that engagement with digital media has substantial associations with adolescents’ mental health symptoms at the population level.”

2. E2 Adolescent Mental Health in the Digital Age: Facts, Fears and Future Directions PubMed Central (author manuscript of Odgers & Jensen, J Child Psychol Psychiatry)

   “The review highlights that most research to date has been correlational, focused on adults versus adolescents, and has generated a mix of often conflicting small positive, negative and null associations.”

3. E3 Adolescent Mental Health in the Digital Age: Facts, Fears and Future Directions PubMed Central (author manuscript of Odgers & Jensen, J Child Psychol Psychiatry)

   “None of the studies reviewed above tested, or were powered to test, whether associations differed by developmental stage.”

4. E4 Let’s Stop Shaming Teens About Social Media Use ASCD (Educational Leadership) · 4 January 2024

   “A large meta-analysis of 226 studies in 2022 involving 275,728 participants led by Jeffrey Hancock estimated the overall association between social media and well-being as indistinguishable from zero.”

The Odgers & Jensen review finds "little evidence" that digital media use has substantial associations with adolescents' mental health at the population level (E1). It describes the research as mostly correlational, with conflicting small positive, negative and null associations (E2). It also notes that the studies it reviewed did not test whether associations differ by developmental stage, so it does not support the claim that the harm is larger in children (E3). A 2022 meta-analysis of 226 studies put the overall association between social media and well-being as indistinguishable from zero (E4). This matches the mostly-false criterion: small or inconsistent associations, so saying social media is known to be harming people "quite a lot" overstates the evidence.