One of the biggest problems facing generative AI in the creative world is not simply opposition to AI. It is that the public is being flooded with examples of its worst possible use.
Low-effort books, images, videos, music, and other content can now be produced in enormous quantities. That material is increasingly described as “AI slop,” and the term is beginning to shape how people think about all AI-assisted creative work.
I think OpenAI has an opportunity to demonstrate the other side of that technology.
My proposal is for OpenAI to create a highly selective creator showcase or publishing imprint for serious human-directed works produced with the assistance of AI.
The emphasis should be on selective.
This should not be a platform where anyone can generate a novel over a weekend and publish it with an OpenAI logo. That would make the existing problem worse. Instead, the purpose would be to establish a visible quality standard for what sustained human-AI creative collaboration can look like.
A novel, game, comic, film, audio drama, or other project submitted to the program might need to demonstrate substantial human creative ownership through things such as original worldbuilding, character development, outlines, revision history, artistic direction, editorial decisions, rejected concepts, research, and repeated iteration.
AI use should be disclosed openly.
The question should not be, “Was AI used?”
The more useful question is, “What did the human creator actually contribute?”
There is an enormous difference between someone asking an AI to “write me a science-fiction novel” and uploading the resulting text, and someone spending months or years developing a fictional universe, designing its civilizations, characters, technology, history, themes, terminology, narrative structure, and individual scenes while using AI as a drafting, editing, brainstorming, and production collaborator.
Right now, the publishing and entertainment industries often struggle to distinguish between those two things.
OpenAI could help establish that distinction.
A serious Creator Imprint could operate around several principles:
- Human creative direction must be substantial and demonstrable.
- AI involvement must be transparently disclosed.
- Works must meet professional editorial and artistic standards.
- Mass-generated content farms should be excluded by design.
- Creators should retain ownership of their underlying intellectual property.
- Copyright and attribution issues should be reviewed carefully before publication.
- Selection should be based on the quality of the finished work, not on how little or how much AI was used.
The program could begin with fiction and eventually expand into comics, animation, games, short films, music, interactive stories, and other forms of entertainment.
There is also a larger cultural opportunity here.
The strongest response to criticism about AI slop is not to deny that AI slop exists. It obviously does.
The stronger response is to produce undeniable counterexamples.
Show people a novel with hundreds of pages of coherent worldbuilding and years of creative development behind it.
Show them a game designed by one or two people who previously could never have afforded the art, writing, coding, and production assistance required to build it.
Show them an animated short made by a filmmaker who had the vision but never had access to a studio-sized production budget.
Show what happens when AI does not replace ambition, but amplifies the ability of an individual creator to pursue it.
I am personally experiencing this distinction while developing a large science-fiction series. AI has been involved extensively in drafting and editing, but the underlying project includes a large human-directed canon: civilizations, military doctrine, technology, characters, chronology, linguistic rules, narrative structure, thematic requirements, scene-level corrections, rejected drafts, and continuous editorial decisions. The AI is part of the production process, but it is not the source of the creative intent.
Projects like that currently occupy an uncomfortable middle ground. Traditional publishers may increasingly reject them simply because generative AI participated in the drafting process, while self-publishing places them beside an enormous volume of low-effort generated material.
There should be a third category:
serious human-created work made possible or substantially expanded through AI collaboration.
OpenAI is unusually positioned to help define that category.
The company would gain high-quality examples of what its tools can enable when used with sustained human direction. Creators would gain a legitimate path for openly acknowledging AI rather than hiding it. Audiences would gain a clearer distinction between automation replacing creativity and technology expanding what an individual creator is capable of producing.
The ultimate goal should not be to prove that AI can make entertainment without people.
It should be to demonstrate something much more interesting:
what people can create when the limits on what one person can realistically attempt begin to change.