Hi @jonesuk,
I was thinking about it; and maybe integrating it in the WordPress comments system (but I always wonder who is really using those built-in comments haha).
You would like to stop which topics? The ones in the chat?
I could add an option to check what the user asks, indeed, but that would slow down the overall request, as the input will need to be check. That’s fine for you?
How about adding the moderation filter to the chat. it would be simple to add this https://platform.openai.com/docs/guides/moderation via the moderation api its free to include and quick with low latency.
As I tried the moderation with mwai but it keeps showing false negatives on simple words which are not in the banned word list.
I have added a quick way to play with it, please check the Settings in the new version 🙂
Hi Jordy that is great, but how can we actually implement it to ensure that users who are chatting in the chatbot are not using for the wrong behaviour. As I can see I can play around with the moderation but sure how to ensure it is added to the chat?
Of course I’ll do it 🙂 It’s just a test for now, to see how it behaves. Do you get good results? Like the “flagged” thing, does it work well depending on your input?
Yes it is good, it doesn’t catch every one but catches the main ones and flags them. great stuff.
