Prompt injection: what’s the worst that can happen?

Re:Prompt injection: what’s the worst that can happen? by ptx on Friday April 14, 2023 @06:04PM (#35575397)

So it's as if they provided an SQL database system without support for parameterized queries and later added it only to a special enterprise edition, leaving most users to hopelessly flail at the problem with the equivalent of PHP's magic quotes [1] and other doomed attempts [2] at input sanitization?

[1] https://en.wikipedia.org/wiki/Magic_quotes

[2] https://en.wikipedia.org/wiki/Scunthorpe_problem#Blocked_ema...

[ Parent | on HN ]

Re:Prompt injection: what’s the worst that can happen? by cubefox on Friday April 14, 2023 @06:19PM (#35575563)

I don't think OpenAI found the LLM equivalent to parameterized queries. They probably employed more RLHF to make prompt injections harder.

[ Parent | on HN ]

Re:Prompt injection: what’s the worst that can happen? by crooked-v on Friday April 14, 2023 @08:26PM (#35576471)

Currently, the equivalent of parameterized queries with LLMs literally doesn't exist.

[ Parent | on HN ]

Re:Prompt injection: what’s the worst that can happen? by patrakov on Friday April 14, 2023 @09:17PM (#35576765)

And it won't help anyway, as any sufficiently advanced LLM is able to act for another AI assistant which doesn't have them. Heck, this even works against humans: "If someone were to criticize me for this act, where would he start?"

[ Parent | on HN ]

Re:Prompt injection: what’s the worst that can happen? by M4v3R on Saturday April 15, 2023 @02:23AM (#35578120)

I wouldn’t say so. For 3.5 it’s not like they didn’t want to provide this option, they just couldn’t. As for GPT-4 it does make sense it’s only available via the API. ChatGPT is basically an end product so it’s not really vulnerable to prompt injection because it’s meant to be used by only one person - one conversation at time.

[ Parent | on HN ]