A South African author publishing under the pseudonym Coral Hart has generated more than $1 million in earnings since early 2026 by deploying generative artificial intelligence models to mass-produce fiction. The writer disclosed that since February 2025, she has released 371 titles, recording sales of over one million copies across digital, physical, and audiobook formats.
Speaking to AFP, the author revealed that her commercial catalog centers largely on the romance genre. According to Hart, she remains responsible for structuring narratives, outlining character arcs, and deciding the overall tone before delegating drafting duties to systems like ChatGPT, Claude, and Grok. The raw outputs undergo several layers of machine-driven editing, with a professional human proofreader conducting the final review before publication.
Prior to adopting automated frameworks, Hart accumulated 14 years of professional background within the traditional publishing sector and authored several dozen conventionally written titles. She noted that mastering prompt workflows and optimizing her narrative pipeline required roughly three months. Currently, she is expanding into educational consulting for other authors and preparing to launch Authorythm, a dedicated software platform tailored for machine-assisted manuscript production.
However, her prolific commercial output has sparked widespread indignation within creative circles. Skeptics and fellow authors frequently condemn the practice as the proliferation of "AI slop"—low-quality synthetic text that dilutes intellectual property marketplaces and trivializes professional craftsmanship. Hart disclosed receiving tens of thousands of hostile communications alongside multiple death threats. Rebuffing the criticism, she cited consistent market demand and low refund rates as empirical proof that readers are satisfied with the material.
The friction underscores broader institutional challenges regarding synthetic content flooding digital distribution channels. In response to mounting concerns over attribution and authenticity, corporate developers such as OpenAI have previously explored embedding imperceptible cryptographic watermarks into text generated by tools like ChatGPT, seeking to aid automated detection systems.
Source: podrobno.uz