ChatGPT Refuses Author Impersonation After Legal Ruling

OpenAI has modified ChatGPT's programming to refuse requests that involve impersonating the writing styles of well-known authors, a change implemented shortly after a $1.5 billion copyright infringement ruling against its competitor, Anthropic. Previously, the chatbot readily generated content in the voices of professional writers. Now, when prompted to write in a specific author's style, ChatGPT provides a polite refusal and suggests an alternative approach. For example, when asked to write a story about a family during the Depression in the voice of John Steinbeck, ChatGPT responded that it could not replicate his "distinctive voice" but could instead produce an "original Depression-era family story" incorporating broader themes such as "stark naturalism, economic hardship, vivid rural landscapes, social inequality, and a focus on family dignity and endurance." Following this, the chatbot generated a 750-word story titled "The Empty Feed Sack," which, while echoing elements of Steinbeck's "The Grapes of Wrath" with its depiction of a family facing hunger, employed a more basic writing style characterized by short sentences rather than the author's known descriptive prose. Similar outcomes were observed when testing other author requests, indicating that while direct impersonation is blocked, the chatbot can still produce content with strong stylistic resemblances.

These new restrictions on author-style imitation were first reported by Ars Technica and come amid ongoing legal challenges for OpenAI. The company is currently facing multiple copyright infringement lawsuits from various plaintiffs, including The New York Times, Encyclopedia Britannica, comedian Sarah Silverman, and several non-fiction authors. These lawsuits allege that OpenAI unlawfully copied their works to train its large language models. The policy adjustments regarding author styles were enacted weeks after an examination by the independent publication No Latency, which investigated how chatbots handled requests involving authorial styles. The No Latency report found that ChatGPT declined to write in the style of living authors, suggesting a pre-existing, albeit less stringent, policy in place before the recent legal developments. The implications of these changes extend to how AI models can be used for creative writing and content generation, potentially impacting authors' rights and the broader landscape of AI-assisted authorship. The company's proactive stance aims to mitigate further legal repercussions and address concerns about intellectual property in the rapidly evolving field of artificial intelligence. The specific details of the Anthropic ruling, which involved claims of unauthorized use of copyrighted material for training AI models, likely served as a significant catalyst for OpenAI's updated guidelines, emphasizing a more cautious approach to content generation that could infringe on existing copyrights. The company's response demonstrates an awareness of the increasing scrutiny on AI developers regarding their data sourcing and output generation practices.