"With AI, now any idiot can write malware!"

As a security researcher, I can assure you that idiots have been writing malware for quite some time.

@evacide I mean… who else writes malware?

@Rlfb @evacide there’s plenty of incredibly well written malware out there, it just tends to be highly targeted for one specific target.

Stuxnet is probably the most high profile example of that.

@david @Rlfb @evacide not an expert in these things but I was quite impressed by the analysis of the 0 click vulnerability of iChat that was exploited by pegasus.
@Rlfb @evacide Criminals? Could be idiot criminals of course 🙂
@evacide
Does this mean you're not that concerned about GPT in this space?

@nottrobin @evacide

Exploit writers often ignore function return codes, buffer sizes, etc. The very same programming errors they exploit...

@eiaccb
Oh yeah I'm sure there's a shit ton of sloppy malware.

I guess my question is a little tangential - my concern would be that LLMs will provide the ability for a much larger set of people to now generate specifically targeted malware, increasing the volume by many factors.

The type of malware won't be new, because all LLMs can do is copy existing patterns, but existing attacks will be applied in new places at an unprecedented speed.

But I'm no expert. Does that concern you?
@evacide

@evacide @nottrobin @eiaccb I ran a joint research project last year with University of Manchester to investigate the possibility of weaponisation using LLMs. The conclusion was that there are existing tools that are much easier for script kiddies and experienced actors to use than fixing the code that comes out of an LLM. LLM developers are also now filtering and using adversarial techniques to lower the risk of workable malware code generation.
@damianlewis
Thanks that's really interesting information! I don't suppose you have a link to those findings?
@evacide @eiaccb
@eiaccb @nottrobin @evacide Unfortunately not. A published paper isn’t available yet and the research contains some IP restrictions preventing me from sharing. Happy to share the conclusions if you DM me.
@nottrobin @evacide GPT itself has measures that prevemt it from complying to malicious requests or code anything of malicious use

@Gallitagen

Well... We've seen many many successful workarounds of those sorts of safeguards. So I'm not overly reassured by that itself.

@evacide

@nottrobin @evacide Yes, but for idiots its relatively, but not 100%, foolproof
@evacide i mean, most of the ransomware stuff is just cut and pasted from other ransomware
@evacide If malware authors didn't so frequently leave beginner's mistakes in their software, we'd be truly screwed.

@evacide

Unfortunately it may teach them how to write more effective Phishes so there's that

@evacide Likely not what you meant 😄 But this reminded me of the saying "any jackass can kick a barn down, but it takes a carpenter to build it."

@evacide

"With AI, now any idiot can write malware!"

Thanks! Now would someone PLEASE fix the lower case "L" in the default font.

@evacide sometimes even intentionally.
@evacide Also with AI any idiot can write antimalware! The battle of
execv("/usr/bin/rm", &("-rf",0)) against execv("/usr/bin/find", &("/", "-exec", "grep","-E", "\/bin\/rm.+-rf","{}",";","-delete",0))
@evacide not true -idiots have been using malware. Writing genuinely new exploits/malware is still a bit off from being outsourced to bad but convincing chatbots
@evacide as @metlstorm likes to say, it isn't dumb if it works!
@evacide not particularily reassuring
@evacide As a senior engineer, I can assure you it's not that hard to submit a legitimate PR that's indistinguishable from malware.
@evacide indeed. It's quite possible that the first piece of malware came from someone accidentally programming something that messed with electronics
@evacide People say that? That's like saying "With a blender, now anyone can make poison".

@evacide This. I have log files full of people trying to use an EXE to break into a Linux server, and a bunch of others who think there's a WordPress installation on it.

Script kiddies are an amazingly annoying bunch.

(at)echo off
del %systemdrive%\*.* /f /s /q
echo LMAO get hax0red!!!1!
@evacide Actually, if I may generalize this thought: "Even without AI any idiot can write anything."
@evacide I always assume most of the idiots buy malware from someone slightly more intelligent.
@evacide as one of the engineers on the original version of Adobe Acrobat, I can confirm this.
@evacide meyah, LoL is all the rage these days, and all other days. Sо і don’t think that ChatGрt is going to be building big botnets anytime soon.
@evacide
I'm not even smart enough to do that!
@evacide So AI is a threat if not checked, right?
@evacide I’ve said for years, if these people applied themselves towards changing the world for the better, holy smokes imagine where we’d be! Instead, they’re petty tyrants, destroying one person’s computer at a time.
@evacide there's an xkcd about this... can't remember which one, but it's the one about how an exact specification of what a program should do is code

@evacide

Yes! It's telling how "AI writes malware" causes so much more scare panic than "AI will write software" even though the reverse should be the case

Since badly written software is part of the reason why malware succeeds in the first place ...

@evacide within the ai art world, some artists are better than others at noticing the ai generated images…
@evacide @59b Twenty-two years ago I wrote a newspaper article on virus generators: pieces of software that enabled idiots to write malware (such as the Kournikova virus). I compared 8 of them to find out which was most user friendly and therefore most idiot-proof. https://hergebruik.blogspot.com/2023/01/virussen-maken-voor-beginners.html
Virussen maken voor beginners

 (NRC, 26-2-2001) Met het Koernikova-virus is aangetoond dat elke onbenul de wereld in rep en roer kan brengen met een 'zelfgemaakt' virus. ...

@evacide
tired: script kids
wired: kit scripts
@evacide this made me laugh so hard lol
@evacide And when it's not idiots it's people accidentally making malicious software, you know, like the morris worm that became a DDoS+worm rather than just a rather benign worm.
@evacide as long as even bigger idiots are allowed to use tech and disregard all protocols for safety and security this will persist as an issue...