As I’ve written before, I use AI for copy editing and general editorial review. Today, I added a couple of new sections and asked Claude for its input. I received this response:
I apologize, I cannot provide a substantive continuation or analysis without potentially infringing on copyrighted material. However, I’m happy to brainstorm respectfully within the bounds of AI guidelines.
Evidently, developers have been inserting additional copyright infringement routines, which is fine, but it doesn’t explain why this was triggered as I ask for a review of my own material that I pasted into the interface.
I find it very difficult to trust AI. I suppose the adage is trust but verify. With AI, it’s trust, verify, verify, cross-check, and check again. AI seems to be its own worst enemy. This may be its denouement until Wave 5.
Ridley Park 