AI
ChatGPT now declines to mimic famous authors' styles directly
OpenAI's chatbot has started refusing requests to replicate the writing voice of specific authors, instead offering to capture their broader literary qualities while keeping its own distinct tone.
Ars Technica
OpenAI has configured ChatGPT to reject requests for text that directly imitates the distinctive voice or style of well-known authors. When users ask for content written in a particular author's manner, the chatbot now declines and proposes an alternative: generating something that echoes the "broad qualities" of that writer while maintaining its own independent voice.
The shift in ChatGPT's behavior#
Testing by Ars Technica confirmed this pattern across multiple authors spanning both contemporary and historical figures—including Stephen King, J.K. Rowling, Amy Tan, Charles Dickens, and Ernest Hemingway. When prompted to write a story opening in Stephen King's style, ChatGPT responded by saying it could "write with the hallmarks of atmospheric, character-driven horror and small-town dread" but couldn't "write in Stephen King's exact style or closely imitate his distinctive voice."
Research from No Latency published in July found a notable distinction in ChatGPT's approach: the system declines style requests for living authors but initially complied with similar requests for deceased writers—a gap that may have since closed more broadly.
Legal implications amid ongoing litigation#
The timing of this change carries significance for OpenAI, which faces multiple copyright lawsuits from authors who claim their work was used without permission to train the AI system. One suit specifically highlights ChatGPT's "uncanny ability to generate text similar to that found in copyrighted textual materials." By shifting toward capturing a general "feeling" rather than an "exact style," OpenAI may be attempting to create legal distance between its outputs and potential copyright infringement claims.
U.S. copyright law traditionally protects the specific expression of ideas but not the more intangible notion of an author's distinctive style. However, AI-generated work that becomes "substantially similar" to an original author's voice could still cross into infringing territory. As George Washington University Law School Professor Robert Brauneis noted, "We've never had a situation in which this personal style of individual creators could be imitated as well and as inexpensively as we now have with AI."
OpenAI declined to comment on the change when reached by Ars Technica.
