[ad_1]
Actually, this is a rather complex topic, and I have been thinking about the safe/good way to implement this for a few days.
It could be preset questions, asked right at the beginning or after a defined number of questions. The answer would be directly used as an input, and we could set if it’s a name, email, or a number, to do some validation and make sure the user doesn’t reply a whole sentence.
The AI models themselves aren’t not able to participate in this, however.
Let me know your thoughts 🙂