Grammarly and the Illusion of Editing by Renowned Experts

An Unexpected Discovery

I was surprised to learn last Friday that I had become an editor for Grammarly. This subscription-based writing assistant recently introduced a feature called "expert review." According to Grammarly, this feature is designed to elevate your writing to the next level through insights from professionals, authors, and subject matter experts.

However, upon closer examination, it turns out that these "insights" do not actually come from leading professionals, nor even from humans. In reality, they are generated by artificial intelligence, and the names of the experts associated with them do not necessarily reflect their opinions. A warning on Grammarly's support page specifies that references to experts are for informational purposes only and do not imply any affiliation or endorsement from them.

Prestigious Names Without Consent

The "expert review" feature caught my attention thanks to an article from The Verge, which reported that this feature was launched last August. Among the "experts" mentioned are famous authors like Stephen King, Neil deGrasse Tyson, and Carl Sagan. My name is also included, along with those of many other tech journalists, including former editors of The Verge like Casey Newton and Joanna Stern, as well as other journalists from Wired, Bloomberg, and The New York Times.

Some expert descriptions contain inaccuracies, such as outdated job titles. These errors could have been corrected if Grammarly had asked these individuals for permission to reference their work. No one asked for my permission to use my name in this way, let alone to compensate me for the expert work that my AI clone was apparently doing on my behalf. An annual subscription to Grammarly costs $144.

Thoughts on the Impact of AI

I have long assumed that one day, AI could take my job. I simply thought someone would warn me when that happened. This situation with Grammarly annoyed me, as it reflects the same sense of entitlement that pervades the web and defines the modern AI industry. At the same time, I felt a morbid curiosity about the quality of these expert reviews, and Grammarly in general.

I regularly ask chatbots for corrections, and I've noticed that their suggestions have significantly improved over the past year. But it would never occur to me to pay for a separate subscription just for a writing assistant—especially knowing that Grammarly is simply reselling tokens from OpenAI and other LLM providers at a markup.

Testing the Feature

Not everyone is ready to copy and paste their Google Doc into a ChatGPT window and ask for a correction. Grammarly claims to have 40 million daily users, and it was time for me to see what they were getting. So, I signed up for a free trial of the company's paid product and got to work.

I used a draft of an article by my colleague Ella Markianos about a protest at OpenAI to test the feature. I clicked on the "expert review" button, one of the twelve tools that Grammarly offers in its paid product, which could generously be described as a ChatGPT prompt turned into a button. In my own editing of Ella's article, I followed the system conveyed through thousands of corrections I have received over the years from my fellow journalists.

We exchanged several times, iterating gradually until we were both satisfied with the result. The "expert review" in Grammarly does not work this way. After clicking the button, a box opened to explain what was happening. A series of expert names appeared before me—each seeming less likely to have ever agreed to this than the last.

There was Shoshana Zuboff, the author who popularized the term "surveillance capitalism," and who has spoken out against the extractive nature of Big Tech. There was Claire Wardle, a misinformation researcher whose work examines how to make communities more resilient to propaganda and hoaxes. And there, floating near the top of the draft, was John Carreyrou, the investigative journalist and bestselling author who exposed Theranos.