Prosocial Design Network
@prosocialdesign.bsky.social
Bridging research and practice toward building healthy online spaces. prosocialdesign.org
Social media has been implicated as a divisive global force — though scholars debate how much blame it really deserves. But a new review of 30 empirical studies finds it can also be used to reduce distrust and animosity across divides.
scholarspace.manoa.hawaii.edu/server/api/c...
scholarspace.manoa.hawaii.edu/server/api/c...
January 5, 2026 at 12:36 PM
Social media has been implicated as a divisive global force — though scholars debate how much blame it really deserves. But a new review of 30 empirical studies finds it can also be used to reduce distrust and animosity across divides.
scholarspace.manoa.hawaii.edu/server/api/c...
scholarspace.manoa.hawaii.edu/server/api/c...
Idea for platforms that want to reduce toxic division & boost engagement: uprank posts that use "connective language."
New research finds that people will spend 25% more time reading posts that use phrases like “this is just my opinion” or “I may be wrong."
mediaengagement.org/research/con...
New research finds that people will spend 25% more time reading posts that use phrases like “this is just my opinion” or “I may be wrong."
mediaengagement.org/research/con...
December 22, 2025 at 3:36 PM
Idea for platforms that want to reduce toxic division & boost engagement: uprank posts that use "connective language."
New research finds that people will spend 25% more time reading posts that use phrases like “this is just my opinion” or “I may be wrong."
mediaengagement.org/research/con...
New research finds that people will spend 25% more time reading posts that use phrases like “this is just my opinion” or “I may be wrong."
mediaengagement.org/research/con...
New on the blog: a conversation with Sabhanaz Rashid Diya asks, whose values are baked into “prosocial” tech design? From content moderation to the limits of exporting the DSA, it looks at why global defaults keep failing the Global Majority. www.prosocialdesign.org/blog/prosoci...
Prosocial Design Values in the Global Majority: A Pro-Social Recap
A conversation with Sabhanaz Rashid Diya, Executive Director of Tech Global Institute
www.prosocialdesign.org
December 18, 2025 at 4:09 PM
New on the blog: a conversation with Sabhanaz Rashid Diya asks, whose values are baked into “prosocial” tech design? From content moderation to the limits of exporting the DSA, it looks at why global defaults keep failing the Global Majority. www.prosocialdesign.org/blog/prosoci...
A dense methods paper, but important reading for researchers studying online behavior and prosocial design interventions. Who we study is as important as how.
osf.io/preprints/so...
osf.io/preprints/so...
December 17, 2025 at 8:30 PM
A dense methods paper, but important reading for researchers studying online behavior and prosocial design interventions. Who we study is as important as how.
osf.io/preprints/so...
osf.io/preprints/so...
Giving users the power to customize their feeds - and using AI to do so - seems to be all the rage. SMT gives rundown of what platforms - including Threads, Youtube, X and Insta - are testing out...
www.socialmediatoday.com/news/threads...
www.socialmediatoday.com/news/threads...
Threads Tests New Way to Manually Guide Feed Algorithms
Just say "dear algo" and submit your request.
www.socialmediatoday.com
December 6, 2025 at 2:48 PM
Giving users the power to customize their feeds - and using AI to do so - seems to be all the rage. SMT gives rundown of what platforms - including Threads, Youtube, X and Insta - are testing out...
www.socialmediatoday.com/news/threads...
www.socialmediatoday.com/news/threads...
Reposted by Prosocial Design Network
Here's a peek at our new NoteTracker Dashboard, a searchable database of all (1.7M) available Community Notes on X.
A key feature of the Dashboard is its ability to display both Visible and Not-yet-visible Community Notes from X. notetracker.socialmedialab.ca
A key feature of the Dashboard is its ability to display both Visible and Not-yet-visible Community Notes from X. notetracker.socialmedialab.ca
November 30, 2025 at 12:09 PM
Here's a peek at our new NoteTracker Dashboard, a searchable database of all (1.7M) available Community Notes on X.
A key feature of the Dashboard is its ability to display both Visible and Not-yet-visible Community Notes from X. notetracker.socialmedialab.ca
A key feature of the Dashboard is its ability to display both Visible and Not-yet-visible Community Notes from X. notetracker.socialmedialab.ca
At PDN we focus on designs that improve outcomes on any online platform, but for digital spaces to truly serve communities prosocial design needs to be part of a platform's core. That's why it's so exciting to see this new platform from @newpublic.org go public and enter beta!
joinroundabout.com
joinroundabout.com
Roundabout
Roundabout is your new online community for local life. Find what you need, connect with neighbors, and take pride in your community.
joinroundabout.com
November 30, 2025 at 2:36 PM
At PDN we focus on designs that improve outcomes on any online platform, but for digital spaces to truly serve communities prosocial design needs to be part of a platform's core. That's why it's so exciting to see this new platform from @newpublic.org go public and enter beta!
joinroundabout.com
joinroundabout.com
New study - and the first field experiment of its kind we're aware of - demonstrates that how much platforms up- or down-rank "antidemocratic attitudes and partisan animosity" in their feeds can have an impact on toxic division.
www.science.org/doi/10.1126/...
www.science.org/doi/10.1126/...
November 28, 2025 at 1:34 PM
New study - and the first field experiment of its kind we're aware of - demonstrates that how much platforms up- or down-rank "antidemocratic attitudes and partisan animosity" in their feeds can have an impact on toxic division.
www.science.org/doi/10.1126/...
www.science.org/doi/10.1126/...
TikTok is taking steps to let users limit how much AI slop they see - 1) adding a preference-slider for AI content and 2) improving methods to detect and water-mark AI content.
www.socialmediatoday.com/news/tiktok-...
www.socialmediatoday.com/news/tiktok-...
TikTok Adds Option To Limit AI Content in Feed
TikTok becomes the latest platform to add an option to limit AI content.
www.socialmediatoday.com
November 28, 2025 at 11:03 AM
TikTok is taking steps to let users limit how much AI slop they see - 1) adding a preference-slider for AI content and 2) improving methods to detect and water-mark AI content.
www.socialmediatoday.com/news/tiktok-...
www.socialmediatoday.com/news/tiktok-...
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...