“Disregard That” Attacks • Hacker News Board

Re: “Disregard That” Attacks

#101

Post by seethishat » Fri, Mar 27, 2026, 3:08 PM UTC

If the main concern is preventing an LLM from taking some action (sending emails, text messages, adding calendar events or making phone calls), can't you just simply not allow the LLM to do that? Don't give it access.

It's not rocket science. If the LLM has no access to do those things, then it can't be tricked into doing those things.

Re: “Disregard That” Attacks

#102

Post by raw_anon_1111 » Fri, Mar 27, 2026, 5:45 PM UTC

gmerc wrote: post #100

raw_anon_1111 wrote: post #76

Earlier quoted context omitted.

And this is another easily solved problem by someone who knows what they are doing… Voice -> speech to text engine -> LLM creates JSON that the orchestrator understands -> JSON -> regular code as the orchestration -> text based response -> text to speech Notice that I am not using the LLM to produce output to the user and if the orchestrator (again regular old code) doesn’t get valid input, its going to error. Sure y…

Could just have used NLP

NLP doesn’t have world knowledge and with one prompt, I can support almost any language. Of course the speech to text engine is specific to the language

Re: “Disregard That” Attacks

#103

Post by chrz » Fri, Mar 27, 2026, 6:19 PM UTC

seethishat wrote: post #101

If the main concern is preventing an LLM from taking some action (sending emails, text messages, adding calendar events or making phone calls), can't you just simply not allow the LLM to do that? Don't give it access. It's not rocket science. If the LLM has no access to do those things, then it can't be tricked into doing those things.

But you want for it to be usefull and do things