Chris Bail

@chrisbail.bsky.social

1178

Followers

309

Following

138

Posts

Duke Professor, Founder of Polarization Lab and Summer Institutes in Computational Social Science. Studying AI, social media, and polarization. Always Breaking the Social Media Prism. Links: www.chrisbail.net sicss.io www.polarizationlab.com

Posts Replies Media

Reposted by

Chris Bail

**Publication Alert!** "An Ideology by Any Other Name," from me & sky-free Chris Johnston in Political Behavior. We investigate ideological identities in the U.S. beyond 'liberal' and 'conservative.' VoR: doi.org/10.1007/s111... Free to read: rdcu.be/dNDiD A brief summary thread:

An Ideology by Any Other Name - Political Behaviordoi.org The terms ‘liberal’ and ‘conservative’ are prominent features of political discourse in the United States, and many citizens choose to identify with one of these ideological labels. Yet, many citizens...

Can LLMs exhibit theory of mind? New in Nature Human Behavior: www.nature.com/articles/s41...

Testing theory of mind in large language models and humans - Nature Human Behaviourwww.nature.com Testing two families of large language models (LLMs) (GPT and LLaMA2) on a battery of measurements spanning different theory of mind abilities, Strachan et al. find that the performance of LLMs can mi...

Reposted by

Chris Bail

Assistant Professorship with Tenure Track in Sociology of Societal Digitalization, University of Zurich

UZH: Assistant Professorship with Tenure Track in Sociology of Societal Digitalizationjobs.uzh.ch The University of Zurich invites applications for an assistant professorship with tenure track in sociology of societal digitalization starting on 1 February 2026. Candidates will be expected to inves...

"we introduce Persona Hub a collection of 1 billion diverse personas automatically curated from web data. These 1 billion personas (~13% of the world's total population), acting as distributed carriers of world knowledge," arxiv.org/abs/2406.20094

Scaling Synthetic Data Creation with 1,000,000,000 Personasarxiv.org We propose a novel persona-driven data synthesis methodology that leverages various perspectives within a large language model (LLM) to create diverse synthetic data. To fully exploit this...

Very cool new paper!

1/ New paper out today in PNAS Nexus! This large-scale study of U.S. book bans started at a summer school 2 yrs ago with @msoligon.bsky.social @isarock.bsky.social Jack LaViolette. This has been such a fun and meaningful project! 🔓Open access paper: academic.oup.com/pnasnexus/ar... 🧵thread below...

Amazing new project that examines how a diverse, cross-national sample of people interact with LLMS.... with data sharing! arxiv.org/abs/2404.16019

The PRISM Alignment Project: What Participatory, Representative...arxiv.org Human feedback plays a central role in the alignment of Large Language Models (LLMs). However, open questions remain about the methods (how), domains (where), people (who) and objectives (to what...

For anyone who has not yet tried running LLMS locally, check out LLAMA 3 7b using Llamafile... Highly recommended! github.com/Mozilla-Ocho...

GitHub - Mozilla-Ocho/llamafile: Distribute and run LLMs with a single file.github.com Distribute and run LLMs with a single file. Contribute to Mozilla-Ocho/llamafile development by creating an account on GitHub.

👑

So excited to announce the new *Iceland* location of the Summer Institutes in Computational Social Science!!! sicss.io/2024/iceland/ thank you to all of the brilliant organizers!

Summer Institute in Computational Social Sciencesicss.io

Is anyone using LLMs to study media bias? Would obviously be challenging for a number of reasons, but also potentially interesting to revisit this long suffering literature with the type of new methods I describe in this new article: www.pnas.org/doi/10.1073/...

Can Generative AI improve social science? | Proceedings of the National Academy of Scienceswww.pnas.org Generative AI that can produce realistic text, images, and other human-like outputs is currently transforming many different industries. Yet it is ...

If LLMs can simulate human survey responses fairly well, will this enable a new type of data fabrication/academic fraud? A question I wish I would have posed in this new article: www.pnas.org/doi/10.1073/...

PNASwww.pnas.org Proceedings of the National Academy of Sciences (PNAS), a peer reviewed journal of the National Academy of Sciences (NAS) - an authoritative source of high-impact, original research that broadly spans...

"Social signals predict contemporary art prices better than visual features, particularly in emerging markets" shorturl.at/BJBNP

Social signals predict contemporary art prices better than visual features, particularly in emerging markets - Scientific Reportsshorturl.at Scientific Reports - Social signals predict contemporary art prices better than visual features, particularly in emerging markets

Interesting new report from Project Liberty reviews attitudes towards AI around the globe. www.projectliberty.io/insight-repo...

Insight report | Project Libertywww.projectliberty.io Project Liberty is building a global alliance for responsible technology and bringing together technologists, academics, policymakers, civil society and citizens to build a safer, healthier tech ecosy...

A typology of successful misinformation interventions: www.nature.com/articles/s41...

Toolbox of individual-level interventions against online misinformation - Nature Human Behaviourwww.nature.com Kozyreva et al. review evidence from individual-level interventions for fighting online misinformation featured in 81 scientific papers. They classify the interventions in nine different types and sum...

Exciting plot twist in the TikTok story: Frank McCourt, who has launched a great new non-profit that works on responsible technology called Project Liberty is now a serious player in the game to acquire the platform: www.semafor.com/article/05/1...

Tech critic McCourt mounts TikTok bidwww.semafor.com The billionaire plans to rebuild TikTok as

Looking for the perfect birthday present for Harrison Butker? @jessicacalarco.bsky.social has you covered!!!!

CAN GENERATIVE AI IMPROVE SOCIAL SCIENCE? I spent the last year trying to answer this question with the help of some very thoughtful reviewers (and editors) for the Proceedings of the National Academy of Sciences (ungated) : www.pnas.org/doi/10.1073/...

It is *vital* not to understate the prevalence of anti-semitism, but we must also be careful not to overstate it as well. Previous reports that 1 in 5 young people deny the Holocaust were probably very inaccurate according to this careful analysis from Pew: shorturl.at/rCHN1 .

Congratulations to @HajYazdiha, for being named a Carnegie Fellow today!!! www.carnegie.org/awards/honor...

Hajar Yazdiha | Carnegie Corporation of New Yorkwww.carnegie.org

Interesting new research on how extremist movements spread across social media platforms: www.nature.com/articles/s44...

Big advance in open-source LLMs: a new approach that helps prevent bad actors from repurposing them for malicious ends-- though the same approach might also enable censorship. Paper here: arxiv.org/abs/2404.126...

SOPHON: Non-Fine-Tunable Learning to Restrain Task Transferability...arxiv.org Instead of building deep learning models from scratch, developers are more and more relying on adapting pre-trained models to their customized tasks. However, powerful pre-trained models may be...

Interesting argument: don't underhype *or* overhype the Russian disinformation effort: www.foreignaffairs.com/russian-fede...

Don’t Hype the Disinformation Threatwww.foreignaffairs.com Underplaying the risk helps foreign propagandists—but so does exaggerating it.

Our new paper in Science Advances outlines standards for research with machine learning, bringing together scientists from many different disciplines to try to identify best practices: www.science.org/doi/10.1126/...

REFORMS: Consensus-based Recommendations for Machine-learning-based Sciencewww.science.org We provide a checklist to improve reporting practices in ML-based science based on a review of best practices and common errors.

Heck yeah Beavers!

Beavers are helping fight climate change, satellite data shows: www.space.com/beaver-dam-n...

Beavers are helping fight climate change, satellite data showswww.space.com Satellites reveal denser green patches of vegetation where beavers were reintroduced.

Congrats to Duncan Watts, who was just elected to the National Academy of Sciences!

Science education is so vital. Check out the amazing work of "Science Journal for Kids" which recently covered our research on using AI chatbots to moderate conflict in sensitive political discussions: www.sciencejournalforkids.org/articles/how... Lots of other content created by a small team.

How can AI make online chats kinder? - Science Journal for Kids and Teenswww.sciencejournalforkids.org + Audio version of article + Blackboard version of article + Links to lessons on political discourse and civil debates

Ever wish your research *did not* replicate? Our study that suggests Americans become *more* polarized when they are exposed to a common enemies (not less) may generalize to Korea and Israel: shorturl.at/dlnCF (original study here: shorturl.at/almT8)

What a beautiful Essay on "Re-wilding" the internet, by Maria Ferrell and @robin.berjon.com : www.noemamag.com/we-need-to-r...