Hi! I’m currently using gpt-4o vision api and testing a few images with it.
And i found out that when i try to upload a celeb image and ask ‘who is this?’ for example, it responds as ‘i can’t figure out who this is’.
But when i try the same thing in chatgpt app, it responds with correct name of the celeb and some description.
Is this some kind of guardrail to limit users from personal identity? if so, why does it work in chatgpt app?
Thank you