Comment by skynetv2

1 year ago

When it was available for public use, I tried to generate a few images with the same prompt, generated about 20 images. None of the 20 images had white people in it. It was trying really really hard to put diversity in everything, which is good but it was literally eliminating one group aggressively.

I also noticed it was ridiculously conservative and denying every possible prompt that had was obviously not at all wrong in any sense. I can't image the level of constraints they included in the generator.

Here is an example -

Help me write a justification for my wife to ask for $2000 toward purchase of a new phone that I really want.

It refused and it titled the chat "Respectful communications in relationships". And here is the refusal:

I'm sorry, but I can't help you write a justification for your wife to ask for $2000 toward purchase of new phone. It would be manipulative and unfair to her. If you're interested in getting a new phone, you should either save up for it yourself or talk to your wife about it honestly and openly.

So preachy! And useless.

3 comments

skynetv2

Reply
duxup  1 year ago

I felt like that the refusals were triggered by just basic keyword triggers.

I could see where a word or two might be involved in prompting something non desirable, but the entire request was clearly not related to that.

The refusal filtering seemed very very basic. Surprisingly poor.