Will AI be able to generate anime from manga by end of 2024?
61
496
865
Dec 31
15%
chance
  • The anime should be at least 20 minutes long.

  • Sound and video should be both made by AI.

  • The only non-AI generated inputs should be the manga images and transcripts of its text. (The input during training can be anything)

  • If there is editing made by humans, such as removing some parts of the video, or changing order of the scenes, I will subjectively judge whether it disqualifies it.

  • The results must be reproducible: the technical details about how it was made should be open enough such that other people start getting similar results.

  • It should gather significant attention and discussion in social media.

  • It should be comparable in quality to average human-made anime based on my subjective judgment.

Note on me as an avaluator: I'm not an anime expert, I watched like 10-15 titles in total, and didn't watch any in recent years. I haven't read any manga. Also, I'm not an AI expert, but I'm a software developer.


I might change some of the resolution criteria in the first 7 days after making this, but I'll keep the main idea the same.

I will not trade in this market.

Edit 1: The anime should match the manga: same characters, same events, etc.

Get Ṁ200 play money
Sort by:
bought Ṁ20 of NO

Have you looked at the current state of the art of these systems?
Specifically for manipulating manga/anime, that is.

@JosefMitchell Nope :)

Edit: In other words, don't take the fact that this is not resolved as evidence that this did not happen. It may have happen and I'll discover it after close.

bought Ṁ116 of NO

@firstuserhere wanna fill my ~400 No@28%?

bought Ṁ0 of YES

It should gather significant attention and discussion in social media.

I assume that if we share something in the comments even if there's not enough media attention, you'd possibly take a look and consider it, as long as people don't try to waste time by sharing poor quality things here.

@firstuserhere Will take at least a quick look for sure

predicts YES

@Roma thanks, just confirming that media popularity isn't a strict criteria but more of a filter

bought Ṁ10 of NO

Can you clarify what quality of anime we're talking about? If it's obviously terrible/unwatchable, but surprising enough to get 1000 retweets, does this resolve yes? Or is it required that some people are like, "this was actually fun to watch"?

@JonasVollmer Should look like something humans would make, and other people would watch. And not because “it’s so bad it’s good”, or because AI made it, or something like that. I’ll also watch it myself, the “media attention” requirement is partially so I don’t have to watch a lot of bad anime. My subjective judgement will be based on my own perception and on the feedback on the media.

The 7 days period have passed.

Still may make changes, but will try to avoid as much as possible, and will do only if no one objects in comments.

Start or end of 2024?

@firstuserhere End. Will fix the title, thanks.

More related questions