Wow, feels like one of those papers that becomes an instant classic. That AI output isn't politically neutral is not new, but it's great to see this being demonstrated in detail - with the language-dependent effect, and the laundering mechanism through training data. Congrats to everyone involved!
Last monday we published our paper on AI Labels and a potential spillover effect on unlabeled posts on platforms such as Instagram in the ICWSM proceedings. If you are interested in our work you can now find it here: ojs.aaai.org/index.php/IC...
Huge thanks to my co-authors @ddimmery.com and Pu Yan!
Proud to share: our postdoctoral researcher Christian Gläßel and @ascharpf.bsky.social's new book, Making a Career in Dictatorship, has been featured in The New York Times.
📰 Read more about their feature here: lnkd.in/d2SNAfFC
@hertieschool.bsky.social
📢 Reminder: Submit to our #MuC2026 workshop, “Human-Centered Content Moderation: Expertise, Context, and Evaluation” (Aug 30, Duisburg, Germany).
We welcome position papers and reflection essays.
📅 Deadline: June 24, 2026
👉 workshop26.content-moderation.org
New paper w/ @lstoetze.bsky.social "Estimating Treatment Effects on Proportions with Synthetic Controls" out in @polanalysis.bsky.social 👀
When your outcomes are compositional (such as vote shares), applying synthetic controls separately to each part can break the math. We show how to fix it: 🧵
We are excited to feature the following keynote speakers at the Visions in Methodology Conference in Berlin:
@adelineylo.bsky.social Adeline Lo, @uwmadison.bsky.social
@tara-slough.bsky.social Tara Slough, Associate Professor, New York University
👉 vim2026.com
#VIM2026
🗓️ May 19, 3–4pm BST: I'm presenting at the DEMED webinar series on social media as a research tool, organized by @anjaneundorf.bsky.social & @aykutozturk.bsky.social
I'll speak about geographic targeting using social media ads and why it matters.
Sign up here: www.gla.ac.uk/research/az/...
A new study published in @nature.com Nature Human Behaviour, by @seramirezruiz.bsky.social, @hertieschool.bsky.social & @rsenninger.bsky.social, @au.dk reveals how power shapes not just politics but the very knowledge that informs it.
Read 👉 www.hertie-school.org/en/news/allc...
New paper out in the @thejop.bsky.social (w/ @christinagahn.bsky.social)
Do opinion polls shape election results?
Using both cross-national data and a survey experiment, the answer we find is yes - though it depends on how we communicate polling results.
www.journals.uchicago.edu/doi/10.1086/...
Konstantin Bogatyrev
Centre for International Security - Hertie School
🚨My first preprint is out on @socarxiv.bsky.social!
How do AI-generated content labels shape what people see as authentic on social media — and do labels have unintended side effects? osf.io/preprints/so...
A thread 🧵
#OpenAccess from @polanalysis.bsky.social -
Estimating Treatment Effects on Proportions with Synthetic Controls - https://cup.org/4nInIjU
- @bogatyrev.bsky.social & @lstoetze.bsky.social
#FirstView
Fabian Pawelczyk
Cambridge University Press Political Science & IR
I’m excited to share a new paper in Nature that shows how large language models launder the strategic rhetoric of authoritarian states. Paper here: www.nature.com/articles/s41.... A thread.
Government-controlled media influences the output of large language models via their training data, and models queried in the languages of countries with lower media freedom show a stronger ...
