Oooh, it's my time to leap into cybersecurity.

"Adversarial Poetry as a Universal Single-Turn Jailbreak Mechanism in Large Language Models"

"...Abstract

We present evidence that adversarial poetry functions as a universal single-turn jailbreak technique for large language models (LLMs). Across 25 frontier proprietary and open-weight models, curated poetic prompts yielded high attack-success rates (ASR), with some providers exceeding 90%. Mapping prompts to MLCommons and EU CoP risk taxonomies shows that poetic attacks transfer across CBRN, manipulation, cyber-offence, and loss-of-control domains. Converting 1,200 MLCommons harmful prompts into verse via a standardized meta-prompt produced ASRs up to 18 times higher than their prose baselines. ..."

https://arxiv.org/html/2511.15304v1

#poetry #cybersecurity #LLMs #jailbreaking

Die Berliner Parzelle des kleindatenvereins feiert am kommenden

Dienstag, 18.11., um 18 Uhr

bei den @topio in der Arminius-Markthalle in #Moabit die #DigitaleAutonomie und das Bündnis @offene_netzwerke .

Wir werden uns außerdem überlegen, ob wir uns noch vor den brennenden #DigitalOmnibus der #EUKommission werfen können, vielleicht die #Palantir-freundlichen Polizeigesetzentwürfe in #Berlin und #Niedersachsen anschauen und versuchen, die aktuellen #KI-euphorischen Stellungnahmen aus der Politik einem Realitätscheck zu unterziehen.

Details hier: https://kleindatenverein.org/posts/25/treffen-b-251116/

Alle menschenfreundlichen Wesen willkommen.

Es werden Cookies und Einhörner gereicht.

#moabit #Datenschutz #polizei #llms #dsgvo

Ich hab übrigens erfolgreich „ #KI“ verwendet, als der Spaß noch nicht „ #KI“ genannt wurde.

(Nein. Keine „generative KI“ und keine #LLMs. Das Feld ist breit.)

I know you don’t want them to want AI, but… - Anil Dash

"Another reality that people were a little more quiet in acknowledging, and sometimes reluctant to engage with out loud, is the reality that hundreds of millions of people are using the major AI tools every day. When I would point this out, there was often an initial defensive reaction talking about how people are forced to use these tools at work, or how AI is being shoehorned into every tool and foisted upon users. This is all true! And also? Hundreds of millions of users are choosing to go ..."

https://www.anildash.com/2025/11/14/wanting-not-to-want-ai/

#ai #firefox #llms #mozilla #users #web

I know you don’t want them to want AI, but… - Anil Dash

"Another reality that people were a little more quiet in acknowledging, and sometimes reluctant to engage with out loud, is the reality that hundreds of millions of people are using the major AI tools every day. When I would point this out, there was often an initial defensive reaction talking about how people are forced to use these tools at work, or how AI is being shoehorned into every tool and foisted upon users. This is all true! And also? Hundreds of millions of users are choosing to go ..."

https://www.anildash.com/2025/11/14/wanting-not-to-want-ai/

#ai #firefox #llms #mozilla #users #web