Cognitive integrity
When AI writes the world
The risks of next‑generation models trained on their own mirror image
August 6, 2025
Updated November 12, 2025
Katri Lindgren