Human Review Teams & Privacy

In the 10/9/26 episode of Author Update, Jonathan and Thomas discussed how Claude reported a user to the police. The woman’s use of AI and her “diary” chats with Claude did not surprise me. What did? The fact that there are human review teams that can view users’ chat transcripts. I feel so PollyAnna, because I “trusted” that privacy was something that was respected—at least for paid-tier users.

For clarification: I do not use Claude to “generate” story text. Every written word is my own. In fact, so far, I’ve only used it to analyze my published stories and to help brainstorm ideas for a a short story I’ve started. For example, What kind of culture, society, etc. might give rise to a world and setting as described in my story? or What are some ways I could inject more hope into the ending? (I write dystopian sci-fi.) I also had hoped to utilize Claude to speed up my research time.

Now, I’m thinking about canceling my Claude subscription and going back to my low-tech research and brainstorming workflows so as to protect my IP.

Before signing up for Claude, I would call upon a trusted colleague or an editor (whose contract specifies privacy to protect my IP) to discuss and brainstorm ideas.

@Thomas Umstattd Jr. @Jonathan Shuerger Any good “privacy” best practices for authors who use or would like to experiment with using AI tools? (If you’ve already covered these in a podcast episode or other resource, please include links.) Thank you, in advance, for your input!