Grammarly has officially suspended its controversial “Expert Review” tool, a high-tier feature designed to provide users with automated feedback mimicking the distinct editorial styles of iconic figures such as Stephen King and Carl Sagan. This sudden retreat follows a class action lawsuit filed by journalist Julia Angwin against Superhuman, the entity identified as Grammarly’s parent organization in the legal filing.

The legal challenge centers on the unauthorized use of names and likenesses to train or prompt AI systems, alleging significant violations of privacy and publicity rights. By offering simulated critiques from living and deceased experts without their consent, the platform waded into a murky legal territory regarding digital replicas and intellectual property.

The feature’s discontinuation highlights a growing tension between generative AI capabilities and the protection of individual creative identity. While leadership expressed a continued belief in the value of expert-simulated feedback, early user reports suggested a discrepancy between marketing and reality.

Reviewers noted that the AI-generated personas often produced generic, uninspired commentary that failed to capture the nuanced brilliance of the writers they were meant to emulate. This case serves as a pivotal moment for the industry, signaling that the era of 'borrowed authority' through unconsented AI personas may face rigorous legal boundaries.

Grammarly has officially suspended its controversial “Expert Review” tool, a high-tier feature designed to provide users with automated feedback mimicking the distinct editorial styles of iconic figures such as Stephen King and Carl Sagan. This sudden retreat follows a class action lawsuit filed by journalist Julia Angwin against Superhuman, the entity identified as Grammarly’s parent organization in the legal filing.

The legal challenge centers on the unauthorized use of names and likenesses to train or prompt AI systems, alleging significant violations of privacy and publicity rights. By offering simulated critiques from living and deceased experts without their consent, the platform waded into a murky legal territory regarding digital replicas and intellectual property.

The feature’s discontinuation highlights a growing tension between generative AI capabilities and the protection of individual creative identity. While leadership expressed a continued belief in the value of expert-simulated feedback, early user reports suggested a discrepancy between marketing and reality.

Reviewers noted that the AI-generated personas often produced generic, uninspired commentary that failed to capture the nuanced brilliance of the writers they were meant to emulate. This case serves as a pivotal moment for the industry, signaling that the era of 'borrowed authority' through unconsented AI personas may face rigorous legal boundaries.