Fabrizio Gilardi
banner
fgilardi.bsky.social
Fabrizio Gilardi
@fgilardi.bsky.social
Professor of political science at the University of Zurich • digital technology, AI & politics • resting bitch face • https://fabriziogilardi.org/
Looking forward to the inaugural @epssnet.bsky.social meeting in Belfast, 18–20 June 2026!

All political scientists encouraged to submit a paper. For work on public policy & administration, consider section I’m co-chairing w/ @asmusletholsen.bsky.social

⏳ 7 Nov 2025
🔗 epssnet.org/belfast-2026...
October 23, 2025 at 8:13 AM
Me: do you know who this is?

15yo: isn’t he the actor who played Obelix?
July 17, 2025 at 7:10 PM
Societal AI Research Has Become Less Interdisciplinary

CS-only teams account for a growing share of work on fairness, safety, and other societal concerns in AI. What does this mean for AI governance and the role of SSH?

New working paper, feedback welcome!

arxiv.org/abs/2506.08738
June 11, 2025 at 3:31 PM
Just published in @pnas.org

@emmahoes.bsky.social and I find that people are far more concerned about AI’s immediate harms than about speculative existential risks. Doomsday talk raises concern about catastrophic outcomes but doesn’t reduce worry about today’s risks

www.pnas.org/doi/10.1073/...
April 17, 2025 at 4:15 PM
April 16, 2025 at 3:37 PM
Call for Papers!

📢 6th Annual Digital Democracy Workshop
🏫 University of Zurich
📅 13-14 November 2025
⏳ Deadline to submit paper abstracts: 31 May 2025

democracy.dsi.uzh.ch/app/uploads/...
April 16, 2025 at 1:41 PM
Curious about the backstory: why would a professor of medicine get a response published to a piece on democratic erosion, asking questions like "What is the rationale for the period chosen?"

www.pnas.org/doi/10.1073/...
March 25, 2025 at 4:52 PM
I really don't like seeing those two numbers together in that sentence 👴🏻
January 29, 2025 at 5:51 PM
January 3, 2025 at 4:39 PM
December 26, 2024 at 6:11 PM
Happy our correspondence is published:

+ Narratives on gen AI: existential risks, immediate risks, acceleration, balanced risks
+ They shape public perceptions & policy
+ We need more research to understand their impact

w/ @argohdes.bsky.social A. Kasirzadeh et al.

www.nature.com/articles/s41...
December 20, 2024 at 10:51 AM
I’m seeing a bat-signal for my template so here it is

fabriziogilardi.org/resources/pa...
December 19, 2024 at 9:42 AM
✨ New LLM paper ✨

> Open-source LLMs perform well in text annotation tasks
> Fine-tuned open-source LLMs outperform zero-shot GPT-4
> Fine-tuning beats few-shot training with modest amount of annotated data

w/ @malizadeh.bsky.social @maelkubli.bsky.social et al.

link.springer.com/article/10.1...
December 18, 2024 at 6:08 PM
it’s like raaaaain on your wedding day...
December 18, 2024 at 5:54 PM
A tricky thing with "AI-generated" labels: this video about NJ drones was labeled as "AI generated" (by TikTok, not the creator) but it seems it's because of a voice effect, not the video itself
December 16, 2024 at 1:17 PM
Stockholm Syndrome
December 8, 2024 at 9:55 PM
helpful analogy to remember the difference between precision and recall
December 5, 2024 at 2:27 PM
suggested criteria for the responsible use of LLMs in academic work

www.nature.com/articles/s42...
November 29, 2024 at 4:31 PM
November 23, 2024 at 1:47 PM
November 22, 2024 at 3:55 PM
November 22, 2024 at 12:16 PM
Disclosure of AI-Generated News Increases Engagement but Does Not Reduce Aversion, Despite Positive Quality Ratings

New version, with @emmahoes.bsky.social et al.

arxiv.org/abs/2409.035...
November 18, 2024 at 8:18 AM
November 16, 2024 at 3:20 PM
November 15, 2024 at 3:13 PM
Swiss political scientists will enjoy this artifact
November 14, 2024 at 8:02 PM