Discussion how to secure your GPTs in different levels with my examples

[Part 1]
One of the most significant concerns I have heard recently about GPTs is that it will expose all your uploaded files or the customized instructions if you know how to talk to them.

THAT’S RIGHT! Since the fundamental point here is AI, nowadays, has been working as a human-like brain, that’s why we need to treat it in the way we - HUMAN - are behaving in daily life.

This means we should teach it how to interact-with and respond-to sabotage or malicious actions, not with technical codes or thingy. (Anyway the highest sake of AI is to help ordinary people without coding-hassle). In the security term, we call it #SocialEngineering

This is my take about it. I made GPT White hat hack using custom GPT:

1 Like


Doesn’t he secure it himself?

True. It should. I must update it. Too much garbage added by default. Was it easy or some trick?

BTW, i spent some time to validate top GPTs. Guess result.