Grammarly has shut down a feature that used artificial intelligence to mimic feedback from real authors, academics, and public figures. The tool, called Expert Review, launched last August and generated critiques styled after individuals ranging from renowned scientists to popular novelists, all without their consent.

The company, Superhuman, stated the feedback was based on publicly available information processed by third-party large language models. A disclaimer noted the named experts were not affiliated with Grammarly, but the move sparked immediate criticism from living writers who found their names attached to AI-generated content. A class action lawsuit is now pending against Superhuman.

Initially, the company offered an opt-out process. However, CEO Shishir Mehrotra announced the feature's full disablement in a LinkedIn post today, stating the tool was intended to help users discover influential perspectives and allow experts to connect with audiences. The statement did little to quell the controversy surrounding the unauthorized use of personal identities.

The decision to pull Expert Review marks a significant retreat for Grammarly as it navigates the complex ethical and legal boundaries of generative AI. The company says it is reassessing the feature's future. This incident underscores the growing tension between rapid AI deployment and the rights of individuals whose public work is used to train these systems.

Source: Engadget