Aleksandra Urman
banner
aurman21.bsky.social
Aleksandra Urman
@aurman21.bsky.social

Senior Research Associate, Social Computing Group, U of Zurich. Computational social/communication scientist

Political science 24%
Communication & Media Studies 24%

Ok that's fair enough:))

To be fair it's Scientific Reports, not main nature (and they do have a bit of a reputation of so to say being less stringent with peer review than they should be :)
Preprint alert! 🥳

How are social bias and CSS interconnected? 🤔

@aytalina.bsky.social, @janabernhard.bsky.social, @valeriehase.bsky.social, and I argue that social bias shapes CSS as a field and as a methodology. Progress in CSS depends on engaging with both dimensions! osf.io/preprints/so...
OSF
osf.io
And now AI slop is taking a serious bite out of SocArXiv.

Thread:
In light of record submission rates and a large volume of AI-generated slop, SocArXiv recently implemented a policy requiring ORCIDs linked in the OSF profile of submitting authors, and narrowing our focus to social science subjects. Today we are taking two more steps:
/1
Just found out about this free and straightforward QR code generator in R:

library(qrcode)
qr_code("add_your_url_here") |>
plot()

That's all! No registration or subscription required.
thierryo.github.io/qrcode/
Generate QRcodes with R
Create static QR codes in R. The content of the QR code is exactly what the user defines. We don't add a redirect URL, making it impossible for us to track the usage of the QR code. This allows to gen...
thierryo.github.io
Who, pray tell, is worshipping the liberal arts? Where is the mythical university alluded to here where STEM and business courses are criticized and somehow disincentivized by those in power?
It’s time to stop worshipping the liberal arts | Letters
Letters: While such institutions have intrinsic value, that doesn’t mean they are entitled to be socially favoured or economically exceptional for ever, says Jianyang Geng
www.theguardian.com

Reposted by Aleksandra Urman

Der Kantonsrat Zürich tanzt völlig aus der Reihe. Der Trend zum Thema Gesichtserkennung im öffentlichen Raum geht in die komplett andere Richtung...dank intensiver Kampagnenarbeit der digitalen Zivilgesellschaft. 8 Städte und Kantone insgesamt haben dies explizit VERBOTEN.

dnip.ch/2025/11/26/d...
Das neue Zürcher Datenschutzgesetz könnte bald verfassungswidrig sein - Das Netz ist politisch
Ein seltsam-kuriose Posse ereignete sich diese Tage im Zürcher Kantonsrat (Minute 2:15-29).
dnip.ch

In this moment people have the concept of the thing in their heads somehow, not necessarily visually, but can't remember any word corresponding to it (it helps for this trick to work ofc that I am in the Swiss context and literally every student speaks at least 2 languages).

Ok it's really dumb. But also in case anyone finds it helpful, the best "trick" I found for the classroom to make sure students snap out of equating language and consciousness or simply thinking is asking them whether they ever had a moment when they forgot some word in all the languages they speak.
can’t fucking catch a breath

make it stop
Is this cheese grater conscious? Many users feel they're talking to a real person. Scientists say it's time to consider whether they're onto something.
can’t fucking catch a breath

make it stop
BREAKING: This is huge news, the EU's equivalent of the 🇺🇸Supreme Court's 2015 Obergefell v. Hodges ruling.

🇪🇺Court of Justice just ruled all 🇪🇺countries must recognise same-sex marriages granted in other member states.

This effectively legalises gay marriage across 🇪🇺
www.reuters.com
After a long time in development, {traktok} #rstats is now finally on CRAN!

Whether you have access to the Research API or just want to scrape some pages, traktok has you covered

jbgruber.github.io/traktok/

Reposted by Aleksandra Urman

X users have found that Grok has now been so tuned to praise Elon Musk that there is no task which it doesn’t say he’s the best at. Grok says Musk is fitter than LeBron and could resurrect from the dead faster than Jesus did in three days.

So childish and funny to see a 50+ year old man do this. 🤣

Reposted by Aleksandra Urman

I've been happily running Linux on my laptops since 2004.

You should try it too! Break away from the horror show of mega tech companies and join the free/libre software community. You'll be so glad you did!

Start by installing Ubuntu, here: documentation.ubuntu.com/desktop/en/l...
Install Ubuntu Desktop
In this tutorial, we’ll download and install Ubuntu Desktop 24.04 LTS on your laptop or PC. If you’ve never installed an operating system, don’t worry: Ubuntu makes the process very easy. You’ll en...
documentation.ubuntu.com
This is one of the worst, most horrifying exposures we have ever seen. It is hard to overstate how bad the things that people are using tools like for are, and it's insane to pretend like these tools are anything other than nonconsensual porn machines

www.404media.co/ai-porn-secr...
Massive Leak Shows Erotic Chatbot Users Turned Women’s Yearbook Pictures Into AI Porn
Chatbot roleplay and image generator platform SecretDesires.ai left cloud storage containers of nearly two million of images and videos exposed, including photos and full names of women from social me...
www.404media.co

And beyond Desheng Hu who led the paper the other authors are @joachimbaumann.bsky.social Elsa Lichtenegger @rcfberg.bsky.social Aniko Hannak (all @scg-uzh.bsky.social ) and Christo Wilson (Northeastern).

While we focus on one specific topic and case study, our overall audit framework generalizes to other high-stakes domains.

The prevance of AI overviews and featured snippets varied significantly across query formulations in terms of question type and sentiment, thus leading to potential inequities in information access.

AI overviews appeared in 84% of search results for the pregnancy and baby care topic (case study - English, US). Almost half of the top-cited sources in them were medium/low credibility.

When an AI overview and a Featured Snippet were present on the same page, they provided contradictory advice in 33% of the cases.

Cautionary notes suggesting a user might want to double check the information with a medical professional were rarely present.

Google AI overviews now reach over 2B users worldwide. But how reliable are they on high stakes topics - for instance, pregnancy and baby care?

We have a new paper - led by Desheng Hu, now accepted at @icwsm.bsky.social - exploring that and finding many issues

Preprint: arxiv.org/abs/2511.12920
🧵👇
Auditing Google's AI Overviews and Featured Snippets: A Case Study on Baby Care and Pregnancy
Google Search increasingly surfaces AI-generated content through features like AI Overviews (AIO) and Featured Snippets (FS), which users frequently rely on despite having no control over their presen...
arxiv.org

Reposted by Aleksandra Urman

NEU: Die deutsche und europäische Rolle für #Russland s Aufrüstung ist deutlich größer als bisher bekannt. Das zeigt unsere Analyse von fast 1 Mio Zolldaten: correctiv.org/aktuelles/ru... @correctiv.org @lilithgrull.bsky.social Dylan Carter #Ukraine
Trotz Sanktionen: Europas Maschinen für Putins Krieg
CORRECTIV-Recherche zeigt: Deutsche und europäische Firmen lieferten zehntausende potentielle Dual-Use-Güter nach Russland.
correctiv.org
Our investigation into Irish Department of Justice use of chatbots. The department hides behind disclaimers while deploying misleading chatbots.

@iccl.bsky.social @abeba.bsky.social @johnnyryan.bsky.social

www.iccl.ie/news/irish-d...
Department of Justice chatbots mislead people seeking information
Irish Department of Justice internalises the AI hype and takes no responsibility for misleading chatbots.
www.iccl.ie

Reposted by Aleksandra Urman

💔 Apartment block destroyed in Ternopil due to Russian attack: two killed
Apartment block destroyed in Ternopil due to Russian attack: two killed – photos, videos
An apartment block has been destroyed in a Russian attack on the city of Ternopil on the morning of 18-19 November. Two people are reported to have been killed.
www.pravda.com.ua

I hope so too. But even then, being an academic doesn't mean having high AI literacy (and being a student who uses GScholar doesn't mean this either)

Reposted by Damian Trilling

Oh "AI-powered" Google scholar search is coming. What could possibly go wrong?! I hate this.

Reposted by Aleksandra Urman

Integrationsdispensation: warum für Migrant*innen strengere soziale Normen gelten - Stefan Manser-Egli, @lutzphilipp.bsky.social

Eine neue Studie zeigt, dass nur von einigen Menschen Integration erwartet wird, während andere nicht diesen Erwartungen unterliegen.

www.defacto.expert/2025/11/17/i...

That's a wild guess, very much biased by all the stories from my friends in different industries about senior people (both, age and position wise) using ChatGPT for very dumb reasons. It somehow ultimately boils down either to the above or "ai is cool and hip we all need to use ai".

I think these are the same people who ask ChatGPT for the "in-depth" analysis of some question they really need to know about. But don't have the time or the will to actually examine it. So you want to feel like you are an expert without actually doing the work for that.