Prof. Arnout van de Rijt (EUI). Source: Arnout van de Rijt.

📢 MZES PUBLIC LECTURE ❗️
📆 Wed, 6 May, 2:30-4:00 pm CET
📍 MZES A 231 and via Zoom

Arnout van de Rijt
@arnoutvanderijt.bsky.social , @eui-eu.bsky.social

will talk on

"Luck and Success in Millions of Life Courses"

Please share and join us! 🙂

👉 www.mzes.uni-mannheim.de/en/news/even...

@fraraffaelli.bsky.social is it today? or when does this take place

denn*

Vielleicht sind die ESS Daten schlicht weg zu alt + du implizierst ja selbst Parteinähe =! Wahlverhalten, oder? Wo sind sie den nun alle hin? Und wie groß ist die Grundgesamtheit der ‘Arbeiter’?

Naja, ich bin mir nicht sicher ob das ausreicht, um die Aussage aus dem Weg zu räumen. Könnten die Zugewinne unter Arbeitern nicht einfach auf die letzten Jahre zurückzuführen sein?

the increase seems substantial to me. what was your prior expectation on the increase?

Who produces hate speech? And how does that matter for content moderation?

We show that across different countries and platforms, a relatively small share of users are responsible for a very large share of hate - overall, 5% write 83-100% of hateful content.

www.cambridge.org/core/journal...

This will be the 3rd edition. If you study pol. behaviour and work on #CEE, this is the right place to present your research. Dates: June 25-26. Discussants include @simonhix.bsky.social & @eliasdinas.bsky.social. Highlight: roundtable on dem. erosion with H. Kriesi, M. Svolik & N. Wunsch.

I assume there won’t be slides or recordings, right?

When you see "AI-generated" on Instagram, does it make you more skeptical – or more trusting of everything without a label?

New research from German Instagram users has answers for policymakers pushing for AI labeling laws.
Good news and a warning👇

Congrats Joff!

Thanks for the support!
@hertieschool.bsky.social
@hertiedatascience.bsky.social and @pku1898.bsky.social :)

n/n This is relevant now: the EU AI Act's labeling requirements enter into force August 2026.

Our findings suggest coverage matters. Partial labeling has trade-offs that policymakers should consider.

Feedback welcome! :)

5/n This matters because full labeling coverage is unlikely in real-world settings.
If some content is labeled and some isn’t, users may update how they interpret unlabeled posts by comparison.

4/n But there’s a catch.
We also find an _implied authenticity effect_: exposure to labeled images slightly increased perceived authenticity of _unlabeled_ content (≈ 1/5 the size of the direct effect).

3/n Main result: Labels work.
Both labeling strategies significantly reduced perceived authenticity of AI-generated images (≈ 0.27 SD).

So far, so good.

2/n We ran a pre-registered survey experiment with a _German_ sample (N = 877).
Participants saw Instagram-style posts with:
• no labels
• a process-based label (“AI-generated”)
• a harm-based label (“Misleading”)

1/n Motivated by new regulations requiring disclosure of AI-generated media, me and my co-authors @ddimmery.com and Pu Yan (Peking University) ask two questions:

1️⃣ Do AI labels reduce perceived authenticity of AI images?
2️⃣ Do labels affect how people judge _unlabeled_ content?

🚨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 🧵

Congrats!

Kann man irgendwo den ganzen Vortrag sehen?

Any data that shows that there are fewer mixed income neighborhoods?

causing gentrification AND social segregation seems contradictory to me. What is the right degree of gentrification that counters soc. segregation?

Congrats!

Very cool work! I will definitely have a closer look into this :)

This seems to be very interesting. But I don’t really understand how you can make this claim based on the two observations from the former point. I’d assume that speeches with more media coverage go along with higher stakes and therefore different vocal styles?

new paper by Sean Westwood:

With current technology, it is impossible to tell whether survey respondents are real or bots. Among other things, makes it easy for bad actors to manipulate outcomes. No good news here for the future of online-based survey research

ich hätte auch interesse an solchem interview!

🚨 Applications Now Open for the EUI PhD Programmes 2026-2027!

📊 Economics |⚖️ Law |📘 History | 🏛 Political and Social Sciences

Join the EUI's 50th PhD cohort!

Apply by 15 January 2026 (14:00 CET) for the academic journey of a lifetime! 👉: eui.eu/phd

#EUIPhD #PhDOpportunity

What happens when you reveal an IG post was made by #AI? Can ranked-choice voting reduce #polarization?
These are the questions tackled by this year’s @hertieschool.bsky.social Data Science Thesis Award winners: Fabian Pawelczyk, Corbin Cerny, & Henry Baker.

👉 www.hertie-school.org/en/datascien...