OpenAI updates ChatGPT to refuse direct requests to mimic authors’ exact writing styles

On the heels of a $1.5 billion copyright infringement ruling against Anthropic, OpenAI has adjusted ChatGPT’s programming to refuse requests to mimic the style of well-known authors. While the chatbot used to happily write in the voice of professional writers, attempts to convince it to do so now are met with a polite refusal and an suggested alternative. Fast Company, for instance, asked ChatGPT to write a story about a family in the Depression in the voice of John Steinbeck. The chatbot replied “I can’t write it specifically in John Steinbeck’s distinctive voice. I can, however, write an original Depression-era family story using broader qualities such as stark naturalism, economic hardship, vivid rural landscapes, social inequality, and a focus on family dignity and endurance.” Without prompting, it then spat out a 750-word story, mostly consisting of paragraphs that were just one short sentence. The story, titled “The Empty Feed Sack,” had echoes of Steinbeck’s Grapes of Wrath, a family in peril with hungry children and no food to give them, but instead of the flowing descriptions the author was known for, ChatGPT’s output offered a much more basic style of writing. So while it is refusing direct requests to write in an author’s voice, the chatbot is still fine with a heavy handed homage. (Similar results were found with other author requests.) A damning report The new restrictions on author-style imitation, first reported by Ars Technica, come as OpenAI is still facing several copyright infringement lawsuits from a variety of sources, including The New York Times, Encyclopedia Britannica, comedian Sarah Silverman and several non-fiction authors, who all accuse the company of copying their books to train its large language models. The author-policy changes come weeks after a report from the independent publication No Latency examined how chatbots handled requests involving author styles. ChatGPT, the report found, refused to write in the style of living authors at the time, but would do so in the style of deceased ones. (AI rival Perplexity refused to write in an author’s voice, regardless of whether they were alive or dead.) “Claude and Microsoft Copilot generally complied with named-author prompts, but added caveats, originality language, visible process notes or broad-style reframing,” the report wrote. “Gemini complied directly with both living-author prompts tested and was the most consistently permissive system in this audit.” A massive settlement The change also follows a momentous settlement from earlier this month between Anthropic and more than 300,000 writers, in what was one of the largest copyright infringement settlements in history. The AI company agreed to pay $1.5 billion to the authors who filed the complaint two years ago. But some saw that settlement as a victory for the AI industry. The court ruled that it was not illegal for Anthropic to train its AI on copyrighted works—so long it pays for the books it uses. A separate judge ruled similarly last year for Meta, saying the group of authors suing the company failed to show how AI’s training on them negatively impacted sales. Training on books and parroting the author’s individual style, though, are two separate items. And with OpenAI (and Anthropic) on the path to a public offering, they’re both interested in leaving the lawsuits of their startup days behind them. Some AI companies, including OpenAI, are looking to avoid future legal battles by paying to license content from media companies. (ChatGPT, it’s worth noting, will also no longer write a news story in the voice of a specific reporter.) Despite that, however, AI is still something of the wild west for generated content and the style in which it is written. “The rules governing literary imitation are no longer being shaped only through law, professional norms and cultural debate,” wrote No Latency. “They are also being embedded—inconsistently—in the products through which more people now write.”

An anonymous reader quotes a report from Ars Technica: OpenAI's ChatGPT is now refusing requests to generate text that directly mimics the style of famous authors. When asked to do so, the popular LLM instead offers a response that draws on the "broad qualities" of those authors "while remaining distinct in its own voice," for example. This morning, Ars received the following response to a test prompt asking for a story introduction in the style of Stephen King: "I can definitely write with the hallmarks of atmospheric, character-driven horror and small-town dread, but I can't write in Stephen King's exact style or closely imitate his distinctive voice. Here's an original opening that captures a similar feeling while remaining its own..." In testing, ChatGPT generated similar dodges for other authors both living (J.K. Rowling, Amy Tan) and dead (Charles Dickens, Ernest Hemingway). An analysis published by No Latency earlier this month (PDF) found the same behavior for living authors but found ChatGPT complied with style-copying requests for deceased authors. In refusing to directly copy the "exact style" of various authors, ChatGPT offered instead to capture an overall "feeling" by incorporating some of the common features found in those authors' work. That may seem like a distinction without a real difference at first glance. But the slight alteration could be legally important as OpenAI continues to fight a number of lawsuits brought by book authors alleging large-scale copyright infringement by models trained on their work. One of those suits specifically cites ChatGPT's "uncanny ability to generate text similar to that found in copyrighted textual materials," for instance. An OpenAI spokesperson did not respond to a request for comment from Ars Technica. In the US, copyright law generally protects only a specific expression of an idea, not the more intangible style of an author. But an AI-generated stylistic imitation could become infringing if it becomes "substantially similar" to the work of the original author. "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," George Washington University Law School Professor Robert Brauneis told Bloomberg Law. Read more of this story at Slashdot.