OpenAI Tightens ChatGPT Rules to Block Mimicking Famous Authors

OpenAI has quietly introduced stricter safety filters to ChatGPT, causing the AI chatbot to reject user requests to write in the distinct styles of popular authors, regardless of whether the writers are living or dead.

The operational shift represents a broader enforcement strategy by the AI firm. Previous testing by tech analysts at No Latency demonstrated that while ChatGPT previously turned down prompts targeting living authors, it routinely fulfilled requests imitating deceased creators. Recent user prompts and independent tests show that these stylistic guardrails now extend universally across major literary figures.

When asked to generate a mystery story about a woman investigating her husband’s sudden disappearance in the prose style of Agatha Christie, ChatGPT declined the task. The system issued a specific disclaimer, noting that because Christie’s works remain protected under copyright, it could not generate text that closely mimics her voice. However, the chatbot offered a workaround, volunteering to write an original mystery that incorporated broader narrative traits typical of her era and genre.

The move reflects growing preemptive risk management across the generative AI sector. Large language models are trained on massive datasets containing published literature to master prose, dialogue, and pacing. While copyright law historically protects specific written expressions rather than abstract concepts or writing styles, technology companies face rising litigation from creative groups such as the Authors Guild over unauthorized data ingestion and competitive output generation.

By embedding system-level refusal triggers around specific author names, AI developers aim to shield themselves from infringement claims. Despite the updated restrictions, ChatGPT continues to write content using general thematic tropes and stylistic characteristics, provided users do not explicitly ask the model to replicate a copyrighted author by name.