Confronting the CEO of the AI company that impersonated me

March 23, 2026·1 hr 16 min·5 clips

Shashir Mahotra apologizes for using Nilay Patel's name in Grammarly's AI feature without consent.

The interview begins with a turn. The host says the plan was a broad Decoder conversation about AI's effect on software, platforms, and creativity, until the Grammarly controversy made it personal. Grammarly had put his name to work. Expert Review showed AI writing suggestions from quote-unquote experts, and reporters found names from The Verge and other outlets in the feature. No one had asked. That makes the conversation less theoretical and more awkward in a useful way: can an AI company lean on legal categories when a real person's name is doing persuasive work for the product? The host keeps pulling the answer back to specifics. He separates trademarks, names, likeness, and copyright, then argues that AI is smearing those lines before the law has caught up. Copyright is the tool everyone knows because YouTube, Instagram, and TikTok already have systems for enforcement and monetization. Creators may reach for that system even when the injury is closer to likeness. A face or name can carry trust, and AI products can make that trust look reusable. The host's skepticism stays dry. He is not looking for a slogan; he wants to know whether the guest learned anything beyond finding the line between legal permission and better judgment. The guest does not hand over a tidy answer. He calls it a creator's Catch-22 and talks about the thin boundary between referencing available work and copying it. Mozart and Bach come up, which gives the defense an old shape: copyright has a long paper trail, while AI likeness fights are still being worked out in real time. The tension stays there. Under the confrontation is the Decoder question: when an AI product borrows authority from a real person, what should count as use?

As heard by us

A pointed Decoder episode about AI impersonation, likeness rights, and the gap between legal permission and product judgment.

Nilay Patel goes into this Decoder interview expecting a wider conversation about AI and software, then the story lands on his doorstep: Grammarly's Expert Review feature had used real writers' names, including his, as AI "experts" without permission.

Read the full review in PlayNext →

Why you'd press play

Your host got impersonated by an AI product. Then he scheduled an interview with the CEO.

Read the full recommendation in PlayNext →

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from Decoder with Nilay Patel