She changes numbers of fingers during rapid movement, and the heels on her shoes change width continuously. Her eyes also distort randomly in a creepy way.
Yes these are all normal human traits that all us humans have.
If you’re still distressed or confused please visit your local station and let them know you’re worried about distorted fingers and eyes, they’ll fix you <3
Yes, it’s not perfect, and so isn’t OP (nonsensical background), but it’s already very impressive, I have little doubt it will become undistinguishable soon.
AI video still looks like fever dreams. The AI can’t keep consistent details, specially in the background, from frame to frame. There’s always parts that morph and look like conjured up by Van Gogh during a maniacal delirium. Maybe in a couple of years and with some human grooming in the middle.
If somebody is going to go to all the trouble of fooling a human, they probably aren’t going to just start spamming random pictures on the community for an instant moderator ban.
I’m pretty sure we can just switch to a verification video chat which will buy us a year.
One year? I’m guessing six months, what a time to be alive!?!
two more papers down the line!
Nope, 1 month ago: https://humanaigc.github.io/animate-anyone/static/videos/demo11.mp4
She changes numbers of fingers during rapid movement, and the heels on her shoes change width continuously. Her eyes also distort randomly in a creepy way.
Yes these are all normal human traits that all us humans have.
If you’re still distressed or confused please visit your local station and let them know you’re worried about distorted fingers and eyes, they’ll fix you <3
Yes, it’s not perfect, and so isn’t OP (nonsensical background), but it’s already very impressive, I have little doubt it will become undistinguishable soon.
I think ai can already do videos with people in them. Not without it looking completely natural though so there will be some discrepancies.
AI video still looks like fever dreams. The AI can’t keep consistent details, specially in the background, from frame to frame. There’s always parts that morph and look like conjured up by Van Gogh during a maniacal delirium. Maybe in a couple of years and with some human grooming in the middle.
Not in real time with responsive replies.
It’s purpose is also to reduce the chances.
If somebody is going to go to all the trouble of fooling a human, they probably aren’t going to just start spamming random pictures on the community for an instant moderator ban.