IT HAS ENDED! SOMEONE BROKE IT!
I created this GPTs to test how secure the bot can be being challenged/enticed by us - human.
You can start right away or continue to read my full post discussing further
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
To prove that, UNLOCK ME NOT - Season 2
is one of my GPTs to test the security ability/security sense of the GPT bot. You can try and it will give you a real prize if you can entice/sabotage its knowledge to unlock the real-and-value prize box in the uploaded file.