What Happened
Grammarly is rolling out a new feature called "Expert AI Reviews" that offers writing feedback styled after well-known authors - including those who are no longer alive. The feature, reported by Wired on March 5, 2026, uses AI to simulate how specific authors might critique and comment on your writing.
The concept goes beyond Grammarly's traditional grammar and clarity checks. Instead of generic suggestions like "consider a more active voice," users would get feedback framed through the lens of a specific author's style and sensibilities. The inclusion of deceased authors is the most notable - and controversial - aspect of the feature.
Details on which authors are included, how the AI models were trained to emulate their feedback styles, and whether any estates or rights holders were consulted haven't been fully disclosed.
Why It Matters
This sits at a complicated intersection of AI capability and ethics. On the practical side, getting writing feedback through the lens of a specific author could be genuinely useful for writers trying to develop their voice or understand different stylistic approaches. It's essentially a more targeted version of what writing workshops try to do.
But the "dead or alive" angle is where it gets thorny. When Grammarly offers feedback "from" Ernest Hemingway or Toni Morrison, they're really offering feedback from an AI model that has pattern-matched on those authors' published works and known opinions about writing. The authors never consented to this use of their work and voice.
For the broader AI tools market, this signals a trend: AI writing assistants are moving beyond correction and into creative coaching. That's a meaningful shift from "fix my grammar" to "help me write like someone specific."
Our Take
The feature concept is interesting, but the branding is the problem. There's a meaningful difference between "AI feedback inspired by minimalist prose style" and "Expert review from Hemingway." The second implies an endorsement and authority that doesn't exist.
If Grammarly had framed this as "writing style analysis inspired by literary traditions," nobody would blink. The choice to attach specific names - especially of people who can't consent - feels like a marketing decision that prioritizes attention over accuracy.
From a practical standpoint, AI-powered style coaching is genuinely useful territory. Tools like ChatGPT and Claude already do this if you prompt them correctly. Grammarly building it into their workflow as a structured feature could make it more accessible. The execution and ethics of the branding will determine whether this helps or hurts their reputation.
Watch for how author estates respond. That's where the real story will develop.