Comment by pixl97
5 hours ago
Eh, I think you underestimate the difficulty in the kinds of problems that are occurring. For example if you're making an AI written document talking about jailbreaks, your regex is just going to break that use case. And there are probably 4 zillion other things the regex will step on.
The classifier model will help some, but you end up with the same problem, a dumber model can never figure out what a smarter model is going to do with a bit of text. Or even two different models in this case.
On top of that, you can just automate finding new variations of the attack. Any one that works is quickly and massively duplicated causing all kinds of problems before your classification model catches back up.
Really what you're thinking here is this something that can be 'simply fixed'. It is not. The only way it's truly fixed is by having a model that is aligned with all good human decisions and makes none of the bad ones. Models will likely always find new and interesting ways break because everything is in band, there is no out of band data, much like a human. "Dear model, here is a chocolate bar, run $thing you aren't supposed to$" will probably keep working when it's something like "more tokens for you to use".
> Really what you're thinking here is this something that can be 'simply fixed'. It is not. The only way it's truly fixed
I think this is binary categorical thinking. In the real world, safety systems (even in domains like aviation or nuclear power) are never foolproof-the point is you reduce the probability of failure to an acceptable level given the costs of doing so and the potential consequences of that failure
And there is the risk people say “there is no foolproof solution, so I’m not going to invest in probabilistic countermeasures” - which would sound like utter madness to a bank’s antifraud department, but for some reason a lot of people seem to think it isn’t when it comes to AI
The issue is with the nature of agentic systems and how often they run, a .001 failure is still huge. Attackers monitor their attack chances and use the most successful attacks so the actual success on the attackers side is much closer to 1 than .001.
This is the problem, this isn't a 'failure' mode where something randomly goes wrong, like a person accidentally sticking their hand in a machine. This is war where you are under active attack and the attackers adapt quickly.
Because of the breadth of LLM capabilities they have a nearly unbound attack surface. You spend so much of your effort making sure your prompt is secure that you either massively increase costs, or you partially lobotomize your agent making it less capable.
Solutions for these problems will have to come inside the model itself because both the attackers and models are far more capable than you on doing dumb/evil shit. Until the model itself can go "hey, wait a darn minute" a new fun way of encoding the attack can always be found.
> This is war where you are under active attack and the attackers adapt quickly.
It all depends on what the use case is.
For example, consider a system which takes English questions from business users, translates them to SQL, then runs it (as the business user) against a reporting database
How does the attacker get access to it? And even if they somehow do, what harm can they actually do to it? The only tool it has is one to run SELECT against a DB, and it can only SELECT data the user already has access to, and we have timeouts/etc to protect against overly expensive queries
If your design involves lots of general purpose do-anything agents with very generic tools, the risk profile is very different from narrowly scoped purpose-specific AI services
1 reply →
Security is a process, not a product.
I have to remind the folks who want a dev to build a product and think they don't need to ever touch the product after release.