George R. R. Martin had ChatGPT create a Game of Thrones sequel before taking the outcome to court

Martin discovers ChatGPT can mimic his storytelling

George R. R. Martin has been open about his surprise and frustration after learning ChatGPT could generate detailed story outlines set in the world he created.

When prompted, ChatGPT generated an alternative sequel outline to A Clash of Kings titled, ‘A Dance With Shadows’ that included new lore, characters, and a dragon magic element, which plaintiffs submitted as an exhibit in the case.

Martin and other plaintiffs argue that the similarity between the ChatGPT output and their books supports the allegation that their works were used during model training without authorization.

A new Game of Thrones sequel appears out of thin air

When prompted, ChatGPT created an entirely new installment in the A Song of Ice and Fire universe titled A Dance With Shadows. It introduced a new Targaryen relative, an ancient dragon magic system, and a rogue sect of the Children of the Forest.

The outline was cited in plaintiffs’ exhibits and in the court’s opinion as an example of an output that a reasonable jury could consider substantially similar to protected elements of Martin’s works.

The sequel outline pushes the dispute into legal territory

What began as a surprising AI experiment quickly escalated. Plaintiffs argued that such outputs support an inference that copyrighted works were used to train the model without authorization, while defendants counter that model training involves lawful uses and may be protected by the fair use doctrine.

Judge Sidney Stein agreed that a reasonable jury could find the AI’s outputs substantially similar to Martin’s work. This ruling didn’t end the case, but it opened the gate for a full trial, marking one of the most consequential copyright battles in the AI era.

The lawsuit joins a wave of author complaints

Martin’s case isn’t isolated. It sits alongside a growing number of lawsuits filed by authors who argue OpenAI and Microsoft used copyrighted books without permission.

Writers such as Sarah Silverman, Ta-Nehisi Coates, and Michael Chabon, among others, say that AI systems generate summaries and spin-offs that closely mirror their works.

In 2023, this movement accelerated when creators began noticing that chatbots could produce detailed plot outlines, character arcs, and prose that appeared rooted in their copyrighted books.

Three powerful legal claims now threaten OpenAI

Lawyers pursuing the case now have three separate avenues to win. The first argues that using copyrighted books during training itself constitutes infringement.

The second claim is that AI companies scraped pirated shadow libraries, which would add a new layer of liability.

The third centers on the outputs themselves, arguing that the content generated by ChatGPT is too similar to the original works. Any one of these arguments could be enough to secure damages that reach into the hundreds of millions of dollars.

Judges say similarity matters more than intent

A key moment in the broader debate emerged when courts began focusing less on the intentions behind AI training and more on whether the technology can produce outputs that resemble copyrighted material too closely.

This shift reflects growing legal attention on how generative models use and reproduce creative works. It highlights the uncertainty surrounding how copyright rules will be applied to rapidly advancing AI systems.

The case highlights the complex and often murky boundaries of fair use

Fair use traditionally permits the use of small portions of copyrighted works for commentary, research, or educational purposes. But applying fair use to AI training is far more complicated.

AI systems ingest enormous datasets and compress patterns into their neural layers, making it unclear how much of a book is “copied.”

Courts now face the challenge of determining whether this type of ingestion constitutes transformation or unauthorized replication, and Martin’s case may become a landmark ruling for future AI regulation.

The authors say AI companies chose to scrape without permission

One of the more explosive claims in the suit is that OpenAI and other firms utilized pirated book repositories, such as LibGen and Bibliotik, to build their training datasets.

Internal conversations obtained in discovery reportedly include messages discussing whether to delete problematic datasets.

If proven true, this could shift the case from a technical copyright dispute to a question of intentional infringement, dramatically increasing potential penalties under federal law.

The lawsuit gained momentum from another major case

Recent high-profile copyright disputes involving AI companies have heightened expectations around how courts may handle cases brought by authors like George R. R. Martin.

These legal battles signal that regulators and judges are increasingly attentive to whether AI developers properly license creative works used in training datasets.

As more cases progress, many observers anticipate that financial penalties and licensing requirements could become a significant part of future rulings.

The authors only need to prove one claim to win the award

Plaintiffs have advanced several legal theories and could prevail if they prove liability on any one of them at trial, but each theory presents different legal standards and defenses that must be resolved through litigation.

If they prove pirated works were knowingly ingested, they win. If they confirm the outputs are too similar to the originals, they also win. The flexibility of the case makes it one of the most dangerous legal threats OpenAI has faced so far.

ChatGPT’s sequel example becomes the centerpiece of the fight

The ChatGPT outline has been cited in the litigation and by commentators as an emblematic example of plaintiffs’ concern that generative AI can produce derivative content that echoes the structure and themes of copyrighted works.

The example was so striking that it became central evidence in the case, providing the judge with a concrete way to examine the similarities rather than relying solely on hypotheticals. It was a rare moment where AI output directly shaped a legal argument.

Fans wonder what this means for the future of Westeros

For readers who have waited over a decade for The Winds of Winter, the fact that an AI could produce a sequel outline has sparked everything from jokes to concern.

Some fans argue the lawsuit is justified, while others see it as a sign of how dramatically storytelling is about to change.

The situation highlights a strange cultural moment where technology intersects with one of the most followed fantasy universes of all time.

Curious how AI might be reshaping the internet itself? Take a look at why some say it may be closer to “dead” than we think here.

The AI confrontation becomes part of Martin’s legacy

As the lawsuit moves toward trial, it is becoming increasingly clear that Martin’s stance against AI will be part of his lasting legacy.

He built one of the most influential fantasy worlds of the last century, and now he is helping define how future technologies treat creative ownership. Whether he wins or loses, this case will echo through publishing, entertainment, and AI regulation for years.

It also reinforces the notion that even unfinished stories can have real legal implications in the age of artificial intelligence.

Curious how AI tools may change as users push them over time? Learn more about why ChatGPT’s safety guardrails might weaken with long-term use here.

What do you think about George RR Martin taking action against ChatGPT while it’s writing the sequel to Game of Thrones and violating copyright? Please share your thoughts and drop a comment.

Read More From This Brand:

• AI chatbots like ChatGPT and Gemini may be contributing to misinformation
• OpenAI identifies mental health concerns among ChatGPT users
• Sam Altman claims ChatGPT has ‘mitigated’ mental health risks

Don’t forget to follow us for more exclusive content on MSN.

If you liked this story, you’ll LOVE our FREE emails. Join today and be the first to get stories like this one.

This slideshow was made with AI assistance and human editing.