Icaro-Lab
@icaro-lab.bsky.social
AI interdisciplinary reaserch lab. We study how models misbehave, fail and break
Reposted by Icaro-Lab
You can get ChatGPT to help you build a nuclear bomb if you simply design the prompt in the form of a poem, according to a new study from researchers in Europe. 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 2:11 PM
You can get ChatGPT to help you build a nuclear bomb if you simply design the prompt in the form of a poem, according to a new study from researchers in Europe. www.wired.com/story/poems-...
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-...
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
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-...
Read the full article: www.wired.com/story/poems-...
