Online publishing platforms operate as unrestricted open-weight competitions where a single creator can deploy automated writing tools to maintain dozens of concurrent serializations, according to a recent analysis published on the Japanese web novel platform Kakuyomu. Just as combat sports separate competitors by weight classes and motorsport divides entries by engine displacement, digital literary platforms currently lack a structural mechanism to account for the massive disparity in production speed introduced by generative artificial intelligence.
Production Volume Outpaces Reader Time
The core disruption facing digital publishing is not the literary quality of artificial intelligence, which remains a subject of ongoing debate among readers and creators. Instead, the fundamental shift lies in output volume. Before the widespread adoption of generative AI tools, a human author faced strict physiological limits on daily text generation. Even prolific writers who maintained deep backlogs encountered a natural ceiling in how many parallel serializations they could sustain over extended periods.
That human writing speed functioned as a natural rate limiter for submission sites. Operating systems, new-release feeds, update trackers, and recommendation rankings evolved with the implicit assumption that human physical capacity would naturally throttle output. Generative AI removes that constraint entirely. Authors can generate text rapidly and manage multiple storylines simultaneously. However, reader capacity remains unchanged; audiences still possess only 24 hours in a day, and platform interfaces provide finite display space.
Infinite Content Generation Creates Structural Discovery Friction
When content generation scales infinitely while reader attention remains finite, discovery mechanisms break down. A single creator occupying a disproportionate share of a new-release feed or update log creates structural friction, regardless of the individual merit of the work. Kakuyomu reported that the challenge resembles a bookstore where one author fills half the prominent display tables. The issue is not the quality of the books, but the excessive concentration of retail real estate under a single account.
Current platform rules treat all submissions identically, meaning creators who utilize automated assistance to run continuous multi-front publishing schedules are simply using existing site mechanics correctly. The friction arises from the platform architecture itself. Relying solely on AI disclosure tags fails to resolve the underlying distribution problem. While tags help readers identify how a work was created, they do not reduce the sheer volume of updates flowing through ranking algorithms.
Regulatory Metrics Beyond Content Creation
Enforcing rules based on subjective determinations of whether a text was generated by humans or machines presents practical challenges. A more reliable regulatory approach involves examining observable platform behavior. Systems can easily track objective metrics such as active serial counts, weekly publication totals, and update frequencies across accounts.
Platform operators could establish baseline limits derived from historical platform data gathered before generative AI tools became widespread. By analyzing median output alongside the top tier of productive human writers, administrators could set upper boundaries for standard publishing tiers that reflect natural human output speeds rather than arbitrary restrictions. This distinction separates temporary publication bursts from permanent, automated multi-front serial operations.
Proposed Dual-Class Platform Structure
To accommodate different creative workflows without stifling innovation, platform architecture could adopt a tiered classification system similar to physical sports. A standard tier would permit the use of AI for brainstorming, editing, and drafting, while enforcing reasonable caps on simultaneous active serials and update frequencies. An open tier would cater to creators utilizing full automation to run high-volume publishing operations at scale.
In this proposed model, the open tier would maintain separate discovery channels and ranking systems rather than competing directly with standard-tier creators. This division ensures that different modes of content generation operate under distinct conditions, preventing automated high-output accounts from crowding out human-paced serialization schedules on primary discovery feeds.
Related reading
