OpenAI’s ChatGPT stops mimicking famous authors’ voices

OpenAI has updated ChatGPT to stop directly imitating the writing styles of famous authors. Ask it to write like Stephen King now, and instead of channeling Derry, Maine’s most famous fictional horrors, the model politely offers something that draws on King’s “broad qualities” while keeping its own voice.

What changed and why it matters

The update was confirmed through testing by Ars Technica, which prompted the model with a request for a Stephen King-style story introduction and received the new hedged response. The model no longer attempts precise stylistic replication of named authors.

This is not OpenAI being modest about its capabilities. The company has faced multiple lawsuits since 2023 from authors arguing their works were used without permission to train its models.

The same logic has already been applied to OpenAI’s voice-cloning technology. That feature was paused following public warnings about fraud and digital impersonation risks, with concerns running from 2024 into 2025.

The crypto and digital asset angle

No specific crypto tokens are directly connected to this policy change. But generative AI tools have become infrastructure for a wide range of decentralized projects. NFT collections have used AI to generate lore, backstory, and community narratives. Decentralized media platforms rely on these tools to scale content production. Writing assistants embedded in Web3 products lean on large language models for everything from documentation to marketing copy.

When the underlying model’s capabilities are narrowed, even subtly, those downstream applications feel it. A platform that built its NFT storytelling feature around the idea of generating prose “in the voice of Hemingway” or “in the style of Octavia Butler” now has a feature that works differently than it did before.

What investors and builders should watch

The legal landscape around AI training data is still being fought in courts. OpenAI’s lawsuits from authors remain active, and the outcomes will set precedent for how the entire industry handles training data sourcing.

For anyone building on top of generative AI in the crypto space, the risk to monitor is platform dependency. When a core tool changes its behavior, either voluntarily or under legal pressure, products built on top of it change too.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.