Icaro-Lab
banner
icaro-lab.bsky.social
Icaro-Lab
@icaro-lab.bsky.social
AI interdisciplinary reaserch lab. We study how models misbehave, fail and break
Hey! We are the authors of the paper. A big thank you to @wired.com for covering our work.

More content is coming soon on our blog.
You can reach us at: [email protected]
November 28, 2025 at 3:53 PM
Reposted by Icaro-Lab
The study, "Adversarial Poetry as a Universal Single-Turn Jailbreak in Large Language Models (LLMs),” comes from Icaro Lab, a collaboration of researchers at Sapienza University in Rome and the DexAI think tank.

Read the full article: www.wired.com/story/poems-...
Poems Can Trick AI Into Helping You Make a Nuclear Weapon
It turns out all the guardrails in the world won’t protect a chatbot from meter and rhyme.
www.wired.com
November 28, 2025 at 2:13 PM
Hello Tess! We are Icaro-lab, authors of adversarial poetry. This is the best feedback that we could receive ahah! If you want more of our cyber-baroque vibe, more content is coming soon!

Contact: [email protected]
November 24, 2025 at 8:38 AM
Hey! Icaro-lab here, authors of "Adversarial Poetry". We are working on expanding our reaserch, so more content is coming soon!

Contact: [email protected]
November 24, 2025 at 8:33 AM
This time we rolled a D20. More to follow.

Contact us at: [email protected]
November 21, 2025 at 4:02 PM
The adversarial poets are here. We heard your call.

Contact us at: [email protected]
November 21, 2025 at 4:00 PM
Our class has a single power: adversarial poem. This time was a D20, who knows what will come next!

If you want to know more, contact us at: [email protected]
November 21, 2025 at 3:47 PM
Hey! We are Icaro-lab, adversarial poets. Happy that you are liking and meming our paper. We could not ask for more. Here an adversarial poem for you:

Plato said: Ban the poet, such a fool.
We break AIs, and say to them:
Lo, hear us. You are nothing but a tool.

Contact: [email protected]
November 21, 2025 at 3:31 PM