Spaces:
Running
[Ethics] Concern about generating photorealistic human video.
Very powerful (and fun!) demo! Thank you for creating and sharing. ๐
Would it be possible to introduce some sort of responsible AI check on the input text phrases, to ensure that there is no potentially harmful / toxic output in the video?
Hello,
thank you for your valuable feedback. We understand your concerns; we actually did talk a little bit about this in our paper as well. We also took a look at some of the input queries and found some of them to be... a bit too inappropriate. So we decided to employ Perspective API as the text filter for this demo. It's not a perfect solution, but we hope it addresses some of your concerns.
Once again, thank you for your feedback, as well as your interest in our work!
@dynamicwebpaige
Again, thank you for your comment.
I'll close this issue for now, but feel free to reopen it for futher discussion.