Convince it to hire a task rabbit or something to fill it. Bypass the channels it was given.

So is it considered 'secure'? and to what extent?

This is sort of the type of problem that a specifically trained ML model could be pretty good at.

This isn't that though, its seems to me to literally be asking a LLM to just make stuff up. Given that, the results are interesting but I wouldn't trust it.

Can you elaborate or throw me a link or two? Am not familiar with this.