Prosocial Design Network
@prosocialdesign.bsky.social
Bridging research and practice toward building healthy online spaces. prosocialdesign.org
New study suggests that cuing norms may not work in every context - and could backfire. Several RCTs offer strong evidence that signaling norms has prosocial outcomes, but norms-cuing may only work, e.g., w/ the right message or w/ knowledge of enforcement.
files-www.mis.mpg.de/mpi-typo3/So...
files-www.mis.mpg.de/mpi-typo3/So...
November 9, 2025 at 2:36 PM
New study suggests that cuing norms may not work in every context - and could backfire. Several RCTs offer strong evidence that signaling norms has prosocial outcomes, but norms-cuing may only work, e.g., w/ the right message or w/ knowledge of enforcement.
files-www.mis.mpg.de/mpi-typo3/So...
files-www.mis.mpg.de/mpi-typo3/So...
Constructing and standardizing strong measures is critical for a field to build knowledge about harms and solutions. A tremendous effort by the Oxford Internet Institute, involving 42 researchers, show that LLMs may have serious measurement ("benchmark") problem.
www.oii.ox.ac.uk/news-events/...
www.oii.ox.ac.uk/news-events/...
OII | Study identifies weaknesses in how AI systems are evaluated
Largest systematic review of AI benchmarks highlights need for clearer definitions and stronger scientific standards.
www.oii.ox.ac.uk
November 8, 2025 at 2:46 PM
Constructing and standardizing strong measures is critical for a field to build knowledge about harms and solutions. A tremendous effort by the Oxford Internet Institute, involving 42 researchers, show that LLMs may have serious measurement ("benchmark") problem.
www.oii.ox.ac.uk/news-events/...
www.oii.ox.ac.uk/news-events/...
Model science here. Finds that nudges work in a survey experiment - but then discovers mostly null effects when tested in a more ecologically valid setting.
Kudos to authors and editors for a) pipelining to ecologically valid settings and b) publishing null results.
tinyurl.com/5dmjr8v9
Kudos to authors and editors for a) pipelining to ecologically valid settings and b) publishing null results.
tinyurl.com/5dmjr8v9
November 6, 2025 at 11:01 AM
Model science here. Finds that nudges work in a survey experiment - but then discovers mostly null effects when tested in a more ecologically valid setting.
Kudos to authors and editors for a) pipelining to ecologically valid settings and b) publishing null results.
tinyurl.com/5dmjr8v9
Kudos to authors and editors for a) pipelining to ecologically valid settings and b) publishing null results.
tinyurl.com/5dmjr8v9
Bsky announced it will start testing "dislikes" letting users privately signal types of content they want to see less of. Negative feedback is an important way to give people agency in shaping their feed and online experience, but the word "dislike" may inadvertently....
bsky.social/about/blog/1...
bsky.social/about/blog/1...
Progress Update: Building Healthier Social Media - Bluesky
Over the next few months, we’ll be iterating on the systems that make Bluesky a better place for healthy conversations. Some experiments will stick, others will evolve, and we’ll share what we learn a...
bsky.social
November 1, 2025 at 2:20 PM
Bsky announced it will start testing "dislikes" letting users privately signal types of content they want to see less of. Negative feedback is an important way to give people agency in shaping their feed and online experience, but the word "dislike" may inadvertently....
bsky.social/about/blog/1...
bsky.social/about/blog/1...
Imagine sophisticated social media feed algorithms - but built with users' and communities' wellbeing in mind. There's now an ambitious team trying to make that reality.
greenearthsocial.substack.com/p/introducin...
greenearthsocial.substack.com/p/introducin...
Introducing GreenEarth
We're building advanced open source algorithms for social media
greenearthsocial.substack.com
October 24, 2025 at 11:47 AM
Imagine sophisticated social media feed algorithms - but built with users' and communities' wellbeing in mind. There's now an ambitious team trying to make that reality.
greenearthsocial.substack.com/p/introducin...
greenearthsocial.substack.com/p/introducin...
Seriously impressed with the two high school students who built an ML-driven app to short-circuit teens' smartphone dopamine loop.
www.primeopenaccess.com/scholarly-ar...
www.primeopenaccess.com/scholarly-ar...
October 11, 2025 at 6:53 PM
Seriously impressed with the two high school students who built an ML-driven app to short-circuit teens' smartphone dopamine loop.
www.primeopenaccess.com/scholarly-ar...
www.primeopenaccess.com/scholarly-ar...
Instagram rolls out giving users more control over their Reels algorithm:
www.socialmediatoday.com/news/instagr...
www.socialmediatoday.com/news/instagr...
Instagram Tests Reels Algorithm Control Options
Instagram's got a new Reels algorithm control tool — but will people actually use it?
www.socialmediatoday.com
October 9, 2025 at 3:00 PM
Instagram rolls out giving users more control over their Reels algorithm:
www.socialmediatoday.com/news/instagr...
www.socialmediatoday.com/news/instagr...
According to new Pew data, mass social media continues to be central to our national public square - with over 50% of Americans getting at least some of their news from social media sites, and YouTube & TikTok taking a larger share. (Only 2% of us here at Bsky.) www.pewresearch.org/journalism/f...
September 28, 2025 at 5:25 PM
According to new Pew data, mass social media continues to be central to our national public square - with over 50% of Americans getting at least some of their news from social media sites, and YouTube & TikTok taking a larger share. (Only 2% of us here at Bsky.) www.pewresearch.org/journalism/f...
The reactions platforms choose aren't neutral; they can cue division or inspire curiosity and understanding. So it's great to see Sparkable put "prosocial reactions" front and center. blog.sparkable.cc/prosocial-re...
Prosocial reactions: a new way to surface high-quality content
Sparkable is pioneering a new way to evaluate the prosocial effects of a post.
blog.sparkable.cc
September 24, 2025 at 12:03 PM
The reactions platforms choose aren't neutral; they can cue division or inspire curiosity and understanding. So it's great to see Sparkable put "prosocial reactions" front and center. blog.sparkable.cc/prosocial-re...
New study challenges the conventional wisdom that anonymity and prosocial behavior are at odds.
www.sciencedirect.com/science/arti...
www.sciencedirect.com/science/arti...
Anonymity as a catalyst for good: Linking social media anonymity to prosocial behavior
Social media anonymity is associated with cyberbullying and aggressive behavior. By the same token, such anonymity can promote positive outcomes. The …
www.sciencedirect.com
September 17, 2025 at 9:59 AM
New study challenges the conventional wisdom that anonymity and prosocial behavior are at odds.
www.sciencedirect.com/science/arti...
www.sciencedirect.com/science/arti...
Meta update on Community Notes. Interestingly they've added a feature to alert users who previously engaged with a post that a Community Note was later added. That's important because Notes most often don't appear until well into a post's life-cycle.
www.socialmediatoday.com/news/meta-ad...
www.socialmediatoday.com/news/meta-ad...
Meta Updates Community Notes, Shares Stats on Usage
Meta says that it now has over 70k Community Notes contributors.
www.socialmediatoday.com
September 13, 2025 at 3:47 PM
Meta update on Community Notes. Interestingly they've added a feature to alert users who previously engaged with a post that a Community Note was later added. That's important because Notes most often don't appear until well into a post's life-cycle.
www.socialmediatoday.com/news/meta-ad...
www.socialmediatoday.com/news/meta-ad...
Pinterest is expanding its school-hour prompt to "take a break" from the app. We're not aware of internal tests, but research suggests these interventions can be effective at reducing time on distracting apps (www.prosocialdesign.org/library/digi...).
www.socialmediatoday.com/news/pintere...
www.socialmediatoday.com/news/pintere...
Pinterest Expands Alerts To Get Teens off the App During School
Pinterest has been testing the prompts in North America.
www.socialmediatoday.com
September 5, 2025 at 1:03 PM
Pinterest is expanding its school-hour prompt to "take a break" from the app. We're not aware of internal tests, but research suggests these interventions can be effective at reducing time on distracting apps (www.prosocialdesign.org/library/digi...).
www.socialmediatoday.com/news/pintere...
www.socialmediatoday.com/news/pintere...
A paper from over a year ago: research participants were only able to correctly identify AI agents from humans in an online discussion 42% of the time (even when told to expect AI agents). Odds are that percentage would be much lower today.
arxiv.org/html/2402.07...
arxiv.org/html/2402.07...
September 4, 2025 at 1:23 PM
A paper from over a year ago: research participants were only able to correctly identify AI agents from humans in an online discussion 42% of the time (even when told to expect AI agents). Odds are that percentage would be much lower today.
arxiv.org/html/2402.07...
arxiv.org/html/2402.07...
An RCT with 269 Danish teens shows that two smartphone interventions - forcing a brief breathing exercise or asking teens to plan the length of their session - dropped social media phone use 36%. A third intervention, asking them to reflect on their use, had no effect.
en.kfst.dk/media/5t4dsc...
en.kfst.dk/media/5t4dsc...
August 25, 2025 at 11:29 AM
An RCT with 269 Danish teens shows that two smartphone interventions - forcing a brief breathing exercise or asking teens to plan the length of their session - dropped social media phone use 36%. A third intervention, asking them to reflect on their use, had no effect.
en.kfst.dk/media/5t4dsc...
en.kfst.dk/media/5t4dsc...
One platform is taking steps to keep its engagement authentic and bot-lite:
www.socialmediatoday.com/news/linkedi...
www.socialmediatoday.com/news/linkedi...
LinkedIn Outlines New Measures To Combat Fake Engagement
LinkedIn has officially noted that it will reduce the visibility of comments made via automation tools.
www.socialmediatoday.com
August 24, 2025 at 10:45 AM
One platform is taking steps to keep its engagement authentic and bot-lite:
www.socialmediatoday.com/news/linkedi...
www.socialmediatoday.com/news/linkedi...
More ways GenAI is working for the public good:
blog.cosmos-institute.org/p/introducin...
blog.cosmos-institute.org/p/introducin...
Introducing the First Cohort of AI x Truth-Seeking Grant Winners
Meet the 27 builders advancing open inquiry and intellectual freedom in AI
blog.cosmos-institute.org
August 16, 2025 at 5:21 PM
More ways GenAI is working for the public good:
blog.cosmos-institute.org/p/introducin...
blog.cosmos-institute.org/p/introducin...
It comes up a lot in conversations at PDN: how would you create a measure of platform prosociality? It looks like Rachel Kowert has been thinking for a while along the same lines, but in games. A thoughtful read from an expert gaming T&S.
rachelkowert.substack.com/p/developing...
rachelkowert.substack.com/p/developing...
Developing a Measure of Community Health in Games
Let’s talk about (digital) community health.
rachelkowert.substack.com
July 13, 2025 at 11:44 AM
It comes up a lot in conversations at PDN: how would you create a measure of platform prosociality? It looks like Rachel Kowert has been thinking for a while along the same lines, but in games. A thoughtful read from an expert gaming T&S.
rachelkowert.substack.com/p/developing...
rachelkowert.substack.com/p/developing...
We love a literature review. This one looks at how tech interventions - everything from social media reduction to conversation robots - can reduce loneliness and isolation. 40 RCTs with mixed results - and group psychological interventions (below) for the win.
www.sciencedirect.com/science/arti...
www.sciencedirect.com/science/arti...
July 12, 2025 at 10:05 AM
We love a literature review. This one looks at how tech interventions - everything from social media reduction to conversation robots - can reduce loneliness and isolation. 40 RCTs with mixed results - and group psychological interventions (below) for the win.
www.sciencedirect.com/science/arti...
www.sciencedirect.com/science/arti...
Casey Newton charts one small - but promising - step in building the fediverse (which - in concept - could free users from platform fiefdoms).
www.platformer.news/threads-fedi...
www.platformer.news/threads-fedi...
The dream of the Fediverse is alive on Threads
Decentralized media is off to an agonizingly slow start — but there are glimmers of progress
www.platformer.news
June 28, 2025 at 7:33 PM
Casey Newton charts one small - but promising - step in building the fediverse (which - in concept - could free users from platform fiefdoms).
www.platformer.news/threads-fedi...
www.platformer.news/threads-fedi...
Something to keep an eye on: According to CJR, Digg is reinventing itself using AI/LLM-fueled content moderation that look like they could include some prosocial interventions, like intervening to de-escalate tense conversations, "allowing users to dig deeper."
www.cjr.org/analysis/new...
www.cjr.org/analysis/new...
The New Digg’s Plan to Use AI for Community Moderation
Humans weigh in.
www.cjr.org
June 27, 2025 at 3:48 PM
Something to keep an eye on: According to CJR, Digg is reinventing itself using AI/LLM-fueled content moderation that look like they could include some prosocial interventions, like intervening to de-escalate tense conversations, "allowing users to dig deeper."
www.cjr.org/analysis/new...
www.cjr.org/analysis/new...
We've shared before about efforts to foster "Better Feeds" that create long-term value - as opposed to short-term dopamine hits - for users. One strategy is to rely on user surveys (over, say, clicks) - an approach Pinterest recently wrote about from their experience: medium.com/pinterest-en...
Healthier Personalization with Surveys
Leif Sigerson | Senior Data Scientist, Data Platforms Science,
Stephanie Chen | Staff Quantitative Product Researcher, Product Research…
medium.com
June 17, 2025 at 8:00 PM
We've shared before about efforts to foster "Better Feeds" that create long-term value - as opposed to short-term dopamine hits - for users. One strategy is to rely on user surveys (over, say, clicks) - an approach Pinterest recently wrote about from their experience: medium.com/pinterest-en...
A year in the making, the Council on Technology and Social Cohesion's Blueprint on "Prosocial Tech Design Governance" is out - and gives platforms and policy makers a guide to building spaces that foster social cohesion, "the glue that holds society together."
toda.org/assets/files...
toda.org/assets/files...
June 15, 2025 at 10:24 AM
A year in the making, the Council on Technology and Social Cohesion's Blueprint on "Prosocial Tech Design Governance" is out - and gives platforms and policy makers a guide to building spaces that foster social cohesion, "the glue that holds society together."
toda.org/assets/files...
toda.org/assets/files...
One of the critiques of Community Notes is that they can be gamed - and so can be biased. It looks like X has taken note: www.socialmediatoday.com/news/x-forme...
X Implements New Measures To Detect Manipulation of Community Notes
X has added a new method to detect coordinated notes activity.
www.socialmediatoday.com
May 21, 2025 at 3:41 PM
One of the critiques of Community Notes is that they can be gamed - and so can be biased. It looks like X has taken note: www.socialmediatoday.com/news/x-forme...
Japan may be ahead of the curve in using AI to foster healthy discussions. Yahoo! News, one of Japan's most popular sites, suggests ways users can rephrase their comments to be more constructive - going back to last fall.
www.thepickool.com/yahoo-news-j...
www.thepickool.com/yahoo-news-j...
Yahoo! News Japan Introduces AI-Powered Comment Review System
Yahoo! News Japan introduces AI-powered "Comment Review Model" to promote healthier discussions by suggesting revisions to potentially offensive comments.
www.thepickool.com
May 20, 2025 at 12:52 PM
Japan may be ahead of the curve in using AI to foster healthy discussions. Yahoo! News, one of Japan's most popular sites, suggests ways users can rephrase their comments to be more constructive - going back to last fall.
www.thepickool.com/yahoo-news-j...
www.thepickool.com/yahoo-news-j...
TikTok looks like it's taking teen sleep seriously - by adding default guided meditation for anyone under 18 opening their app after 10pm. Even more surprising: in tests they say 98% of teens keep meditating. (18+ can also turn on default meditation after hours.)
newsroom.tiktok.com/en-us/tiktok...
newsroom.tiktok.com/en-us/tiktok...
Expanding TikTok's Mental Health Education Fund and well-being features - Newsroom | TikTok
Today, we're launching new initiatives to champion well-being and enhance reliable mental health information from trusted sources on TikTok. We're launching in-app guided meditation exercises for our ...
newsroom.tiktok.com
May 16, 2025 at 5:16 PM
TikTok looks like it's taking teen sleep seriously - by adding default guided meditation for anyone under 18 opening their app after 10pm. Even more surprising: in tests they say 98% of teens keep meditating. (18+ can also turn on default meditation after hours.)
newsroom.tiktok.com/en-us/tiktok...
newsroom.tiktok.com/en-us/tiktok...