attack on maik.io https://maik.io/en/tags/attack/ Recent content in attack on maik.io maik.io en Fri, 07 Nov 2025 19:57:28 +0000 92 % https://maik.io/en/notes/2025-11-07-92-prozent/ Fri, 07 Nov 2025 19:57:28 +0000 https://maik.io/en/notes/2025-11-07-92-prozent/ <p>Cisco&rsquo;s study &ldquo;<a href="https://arxiv.org/pdf/2511.03247" target="_blank" >Death by a Thousand Prompts: Open Model Vulnerability Analysis</a>&rdquo; shows that so‑called multi‑turn jailbreaks against open AI models (Open-Weight-LLMs) succeed in up to 92% of cases, revealing serious security vulnerabilities in their architecture.</p> <p>The text was automatically translated from German into English. The German quotations were also translated in sense.</p> Cisco’s study “Death by a Thousand Prompts: Open Model Vulnerability Analysis” shows that so‑called multi‑turn jailbreaks against open AI models (Open-Weight-LLMs) succeed in up to 92% of cases, revealing serious security vulnerabilities in their architecture.

The text was automatically translated from German into English. The German quotations were also translated in sense.

]]>

Vielen Dank fürs Lesen! Wenn du Lust auf noch mehr Gedanken, Updates und ab und zu einen Blick hinter die Kulissen hast, folge mir doch gern auf Mastodon oder Instagram.

Hast du Fragen oder Feedback? Schreib mir gern eine E-Mail.

]]> 250 manipulated documents https://maik.io/en/notes/2025-10-16-250-manipulierte-dokumente/ Thu, 16 Oct 2025 04:24:20 +0000 https://maik.io/en/notes/2025-10-16-250-manipulierte-dokumente/ <p>The study &ldquo;<a href="https://arxiv.org/abs/2510.07192" title="Poisoning Attacks on LLMs Require a Near-constant Number of Poison Samples" target="_blank" >Poisoning Attacks on LLMs Require a Near-constant Number of Poison Samples</a>&rdquo; shows that as few as about 250 manipulated documents are enough to poison even large language models with false information.</p> <p>With comparatively little effort, propaganda can be injected or faulty code generated when AI is used in software development.</p> <p>This turns artificial intelligence into the largest black box in IT.</p> The study “Poisoning Attacks on LLMs Require a Near-constant Number of Poison Samples” shows that as few as about 250 manipulated documents are enough to poison even large language models with false information.

With comparatively little effort, propaganda can be injected or faulty code generated when AI is used in software development.

This turns artificial intelligence into the largest black box in IT.

The text was automatically translated from German into English. The German quotations were also translated in sense.

]]>

Vielen Dank fürs Lesen! Wenn du Lust auf noch mehr Gedanken, Updates und ab und zu einen Blick hinter die Kulissen hast, folge mir doch gern auf Mastodon oder Instagram.

Hast du Fragen oder Feedback? Schreib mir gern eine E-Mail.

]]>