Things That Will Lose Value / Rights Management Will Become Production Capability Itself / How to Make AI Dramas [April 2026 Edition]
I have released Episode 1 of the AI drama produced with Seedance 2.0.
Having used video generation AI for over three years since Runway Gen-1, Seedance 2.0 is the momentous model with which I was 'able to produce exactly as I intended for the first time.'
Instead of connecting good-looking clips through massive gacha-like generation, I was able to control and plan the video production according to the scenario. However, the difficulty of production has become terrifyingly high, so it is no longer a task that can be done on the side...
What is important here is that the lower limit of video quality has risen all at once.
Until now, low-budget works were obvious the moment you saw them, but from now on, a massive amount of video that looks high-quality even if it is low-budget will emerge. In other words, visual quality and impact alone will no longer function as strong differentiators for viewers or clients.
First, I will introduce the behind-the-scenes of how it was produced.
How to produce Episode 1
I explain the production method in detail in the following video.
It is nearly an hour long, but I also introduce the prompt structure optimized for Seedance 2.0, so please use it as a reference. *This is a re-recorded version of the live stream '1st MicroEpic Club Online Meeting' held on the 25th (Sat) of this month. It only covers the production-related parts, so the content is slightly different.
Playback time: 57 minutes 36 seconds
This is the archive of the live stream for participants. Please enter the viewing password announced on the day. *This includes production time, costs, issues with Seedance 2.0, and evaluations of AI aggregators. It is not included in the public video (re-recorded version) above.
Playback time: 65 minutes
Viewing period: Until 11:59 PM on Friday, May 29, 2026.
In traditional video production, the ability to shoot itself was the value. Being able to gather actors, go on location, build sets, create lighting, and composite CG. These were all expertise, costs, and barriers to entry.
However, in AI video, much of that is compressed. Of course, it does not disappear completely. Acting guidance, physical art, and the serendipity that can only be obtained on-site will continue to have value in the future. However, for short- and medium-length video content watched online, the areas where 'video that would have been impossible without shooting' is replaced will expand rapidly.
Things that will lose value from now on
First, what will lose value is one-off visual impact. This is because many people will use the same model, write prompts with the same cinematic vocabulary, and be able to output the same kind of high-quality video.
The screen is beautiful, but it leaves no memory. Such videos will increase. What is scary is not low-quality failures. It is video that is high-quality but leaves nothing behind. Isn't this the area that creators should be wary of?
I think everyone senses the asymmetry that the more AI slop floods the market, the higher the scarcity premium for human-made works will rise. As AI content overflows on every platform, the scarcity of human craftsmanship will increase in value.
This indicates that the market for human production will not disappear in parallel with the spread of AI, but will instead polarize. The lower end will become completely commoditized, while the upper end will conversely become premium. In other words, the middle will disappear.
Rights management will become production capability itself
From now on, rights management will not be administrative work, but the core of creativity.
Which model was used? Which reference images were used? What is the status of the rights for the reference materials? Is it too similar to a real person? Is it too close to the designs of existing movies or games? Is it necessary to disclose the use of AI upon release?
A production system that can record and explain this information will be necessary.
This is not just a story for major studios.
Individual creators and small teams will not be able to avoid this in the future either. In fact, individuals are more directly affected by account suspensions and rights claims from tools (services).
Future video creators must not only make good videos, but also make videos that can be released with peace of mind.
In that sense, rights design will become a competitive advantage.
Failure logs are the greatest production asset
What is important in AI video production is the record of failures.
Under what conditions do characters swap? Which reference images are easily blocked? Which prompts induce excessive effects? Which scenes are prone to breaking down? Which models succeed and which models fail?
Such failure logs cannot be easily bought from the outside.
This is because they can only be accumulated within one's own work and one's own workflow.
Many people will be able to make videos.
However, the number of people who can continuously operate video production will be limited.
I believe that difference is the true creative competitiveness that will remain in the AI era.
Updated: April 28, 2026 (Tue) / Published: April 28, 2026 (Tue)