Grammarly Doing Some Creepy AI Stuff

Key Takeaways

• Grammarly used Stephen King's writings to train AI
• No permission obtained from author or estate
• Raises copyright infringement concerns for AI models
• Potential right‑of‑publicity violation over name usage
• Signals broader regulatory scrutiny of AI training data

Pulse Analysis

Grammarly’s latest AI feature promises users a “Stephen King‑style” writing coach, built by feeding thousands of the author’s novels and short stories into a large language model. Leveraging the same data‑hungry techniques that power ChatGPT, the tool extracts stylistic cues—sentence rhythm, vocabulary, and narrative pacing—to suggest edits that echo King’s trademark voice. While such personalization can boost user engagement, the underlying data pipeline often operates in a gray area, pulling publicly available text from the internet without explicit licensing. This practice reflects a broader industry push to monetize recognizable literary tones.

The lack of permission raises immediate legal red flags. Copyright law protects the expressive content of King’s works, and reproducing substantial portions in a trained model may constitute infringement, especially if the output can be traced back to the original text. Moreover, the right of publicity shields an author’s name and persona from commercial exploitation, meaning Grammarly’s branding of a “Stephen King analyzer” could violate those rights. Recent lawsuits against AI firms—such as the Authors Guild case against OpenAI—demonstrate that courts are willing to scrutinize unlicensed training data, potentially imposing damages and injunctions.

Grammarly’s misstep is prompting a reassessment of data‑use policies across the AI sector. Companies are now exploring licensing agreements with rights holders, developing synthetic datasets, or implementing robust attribution frameworks to mitigate risk. Regulators in the EU and US are also drafting guidelines that could require transparency about training sources and consent mechanisms. For businesses, the lesson is clear: investing in legal vetting and ethical sourcing is no longer optional but a prerequisite for sustainable AI product development. Properly managed, AI‑enhanced writing tools can still thrive without infringing on creators’ rights.