AI Infosec

743 readers
1 users here now

Infosec news and articles related to AI.

founded 1 year ago
MODERATORS
[email protected]
listing: all - local
1
41
Stealing everything you’ve ever typed or viewed on your own Windows PC is now possible with two lines of code — inside the Copilot+ Recall disaster. (doublepulsar.com)
submitted 3 weeks ago by [email protected] to c/[email protected]
0 comments fedilink
2
3
Anyscale addresses critical vulnerability on Ray framework — but thousands were still exposed (venturebeat.com)
submitted 2 months ago by [email protected] to c/[email protected]
0 comments fedilink
3
35
AI hallucinates software packages and devs download them – even if potentially poisoned with malware (www.theregister.com)
submitted 3 months ago by [email protected] to c/[email protected]
3 comments fedilink
4
3
Why Are Large AI Models Being Red Teamed? (spectrum.ieee.org)
submitted 3 months ago by [email protected] to c/[email protected]
1 comments fedilink
5
6
How 'sleeper agent' AI assistants can sabotage code (www.theregister.com)
submitted 5 months ago by [email protected] to c/[email protected]
0 comments fedilink
6
29
NIST: If someone's trying to sell you some secure AI, it's snake oil (www.theregister.com)
submitted 5 months ago by [email protected] to c/[email protected]
1 comments fedilink
7
3
Boffins devise 'universal backdoor' for image models to cause AI hallucinations (www.theregister.com)
submitted 6 months ago by [email protected] to c/[email protected]
0 comments fedilink
8
2
LLM Finetuning Risks (llm-tuning-safety.github.io)
submitted 8 months ago by [email protected] to c/[email protected]
0 comments fedilink
9
4
Are Local LLMs Useful in Incident Response? - SANS Internet Storm Center (isc.sans.edu)
submitted 9 months ago by [email protected] to c/[email protected]
0 comments fedilink
10
25
Microsoft Bing Chat spotted pushing malware via bad ads (www.theregister.com)
submitted 9 months ago by [email protected] to c/[email protected]
0 comments fedilink
11
5
New AI Beats DeepMind’s AlphaGo Variants 97% Of The Time! (www.youtube.com)
submitted 9 months ago by [email protected] to c/[email protected]
0 comments fedilink
12
4
Identifying AI-generated images with SynthID (www.deepmind.com)
submitted 10 months ago by [email protected] to c/[email protected]
0 comments fedilink
13
6
Thinking about the security of AI systems (www.ncsc.gov.uk)
submitted 10 months ago by [email protected] to c/[email protected]
0 comments fedilink
14
6
GitHub - google/model-transparency (github.com)
submitted 10 months ago by [email protected] to c/[email protected]
0 comments fedilink
15
2
disinformation videos on AI ? (infosec.pub)
submitted 10 months ago by [email protected] to c/[email protected]
10 comments fedilink
 
 

Hi all,

Had a small chat on #AI with somebody yesterday, when this video came up: "10 Things They're NOT Telling You About The New AI" (*)

What strikes me the most on this video is not the message, but the way it is brought. It has all the prints of #disinformation over it, .. especially as it is coming from a youtube-channel that does not even post a name or a person.

Does anybody know this organisation and who is behind it?

Is this "you are all going to lose your job of AI and that's all due to " message new? What is the goal behind this?

(Sorry to post this message here. I have been looking for a lenny/kbin forum on disinformation, but did not find it, so I guess it is most relevant here)

(*) https://www.youtube.com/watch?v=qxbpTyeDZp0

16
8
Universal and Transferable Attacks on Aligned Language Models (llm-attacks.org)
submitted 10 months ago by [email protected] to c/[email protected]
0 comments fedilink
17
13
OWASP Top 10 for LLMs (v1.0) (owasp.org)
submitted 11 months ago by [email protected] to c/[email protected]
1 comments fedilink
18
6
Cybercriminals train AI chatbots for phishing, malware attacks (www.bleepingcomputer.com)
submitted 11 months ago by [email protected] to c/[email protected]
0 comments fedilink
19
-3
GPT Malware Creation (infosec.pub)
submitted 11 months ago by [email protected] to c/[email protected]
4 comments fedilink
 
 

Anyone else getting tired of all the click bait articles regarding PoisonGPT, WormGPT, etc without them ever providing any sort of evidence to back up their claims?

They’re always talking about how the models are so good and can write malware but damn near every GPT model I’ve seen can barely write basic code - no shot it’s writing actually valuable malware, not to mention FUD malware as some are claiming.

Thoughts?

20
7
Adversarial suffixes that circumvent the alignment of open source LLMS, ChatGPT, Claude, Bard, and LLaMA-2 (twitter.com)
submitted 11 months ago by [email protected] to c/[email protected]
0 comments fedilink
21
8
(Ab)using Images and Sounds for Indirect Instruction Injection in Multi-Modal LLMs (twitter.com)
submitted 11 months ago by [email protected] to c/[email protected]
0 comments fedilink
22
7
Intro to ML Safety - Free course (course.mlsafety.org)
submitted 11 months ago by [email protected] to c/[email protected]
0 comments fedilink
23
8
A framework to securely use LLMs in companies - Part 1: Overview of Risks (boringappsec.substack.com)
submitted 11 months ago by [email protected] to c/[email protected]
0 comments fedilink
24
3
Army looking at the possibility of 'AI BOMs' (defensescoop.com)
submitted 11 months ago by [email protected] to c/[email protected]
0 comments fedilink
25
7
Impact of remote-code execution vulnerability in LangChain (ntietz.com)
submitted 11 months ago by [email protected] to c/[email protected]
0 comments fedilink
view more: next ›