Login

Recent searches

No recent searches

Search options

Not available on mastodon.xyz.
mastodon.xyz is one of the many independent Mastodon servers you can use to participate in the fediverse.
A Mastodon instance, open to everyone, but mainly English and French speaking.

Administered by:

Server stats:

817
active users

mastodon.xyz: AboutProfiles directoryPrivacy policy

Mastodon: AboutGet the appKeyboard shortcutsView source codev4.3.4

#aisecurity

1 post1 participant0 posts today
Continued thread
Public *
nonlinear

Solutions? Allow-lists (members can only connect with approved third-parties), app transparency (good luck with that), certifications (or, social allow-lists).

Model Context Protocol adoption will usher an ecology of highly specialized AI tools, and that's a good thing. But can we have it with privacy?

Thoughts?

#AIsecurity#TechEthics#DataProtection
Public
Yetkin Degirmenci 🍫

The Power of Words: Prompt Engineering and Jailbreaks

"Think of it like this: in social engineering, using the right words can open doors, build trust, and unlock information. Similarly, with LLMs, which are trained on vast amounts of human language, choosing the right words in your prompts is key to “opening the door” to clear, insightful, and truly valuable answers."
#AI #PromptEngineering #LLM #AICommunity #AISecurity #AIRedTeaming #AIJailBreaks

medium.com/@yetkind/the-power-

Medium · The Power of Words: Prompt Engineering and JailbreaksBy Yetkin Degirmenci
Public
OWASP Foundation

Ready to Secure AI Systems? Join Our 3-Day Hands-On Training at OWASP Global AppSec EU 2025!

Dive into AI/ML Whiteboard Hacking with expert Sebastien Deleersnyder from May 26-28, 2025 in Barcelona.

Designed for AI engineers, developers, architects, and security professionals, this intermediate-level training will equip you with practical skills to identify AI-specific threats.

owasp.glueup.com/event/123983/

#AppSecEU#OWASP#ThreatModeling
Public
Michał "rysiek" Woźniak · 🇺🇦

I am reading up on abliterations:
huggingface.co/blog/mlabonne/a

Still trying to wrap my head around the consequences of this. But...

...I kinda feel like abliterations have implications also for prompt injections?

As in, it feels like abliterations could mean that it is simply impossible to secure an LLM from prompt injection?

I'm sure I am misunderstanding stuff here. Anyone any input on this?

huggingface.coUncensor any LLM with abliterationA Blog post by Maxime Labonne on Hugging Face
#AI#AISecurity
Public
Norobiik @Norobiik@noc.social

The vulnerability abused long-term conversation memory. Memory with #ChatGPT stores information from previous conversations and uses it as context in all future conversations.

The researcher demonstrated how he could trick ChatGPT into believing a targeted user was 102 years old, lived in the Matrix, and insisted Earth was flat. #AI #AISecurity #GenAI

#Hacker plants false memories in ChatGPT to steal user data in perpetuity
arstechnica.com/security/2024/

Ars Technica · Hacker plants false memories in ChatGPT to steal user data in perpetuityBy Dan Goodin
Public
23Ro

Currently working on a non-profit in #AISecurity ( oais.is ) with some critical work and are looking for hardware sponsoring. Primarily we are looking for Notebooks that we can run Linux on and burner phones.

I've tried to reach out to #tuxedo computers without feedback. Does anyone have a lead on linux focused hardware shops that offer sponsoring for non-profits?

#ai #alignment #security #cybersecurity #linux #linuxnotebooks #notebook #hardware #sponsoring #non-profit #ngo

OAISISOAISISCreating Trust.
ExploreLive feeds
About