Grammarly has disabled an artificial intelligence feature that imitated the voices of well-known writers after strong criticism and legal complaints from those whose names appeared in the tool.
The feature, called Expert Review, generated editing suggestions “inspired by” public figures such as Stephen King and Carl Sagan. It presented feedback as if those figures were reviewing a user’s writing. The tool ran through systems developed by Superhuman, the technology firm behind Grammarly’s recent AI expansion.
Several writers objected after discovering their names attached to advice they never gave. Investigative journalist Julia Angwin filed a class-action lawsuit in the U.S. District Court for the Southern District of New York, arguing that the companies used professional identities without permission to promote a paid service.
Angwin said she was surprised to see her name linked to editing guidance generated by a machine. The lawsuit claims the companies built AI “personas” using the identities of hundreds of writers and academics.
Following the backlash, Superhuman chief executive Shishir Mehrotra said the company would remove the feature and rethink its approach. In a public statement, he acknowledged that the system had misrepresented the voices of experts.
Grammarly launched in 2009 as a digital writing assistant and began adding generative AI tools last year. The controversy highlights a growing debate over how technology companies should use the identities and published work of authors when developing artificial intelligence systems.
Related Readings:
