Using AI to restore audio?

I think this ties into the general debate around the use of AI in the process of restoration. I happen to fall in the camp that it can be a valuable tool if properly used. I think it's more than acceptable for media that can't really be restored by other means. I would rather have a reconstruction than nothing at all. In the case you described it would be even easier for me to support, since the AI would just be expanding upon what English audio is there (Instead of restoring fully lost audio). I would rather deal with artifacts of the AI process than a voice over.