Alignment Forum3 min read
Neural nets are exceptionally vulnerable to goal self-hack
Suppose we have an AI CEO, ruthlessly devoted to maximising shareholder value. Fortunately, we haven't been completely stupid, and have given it a rich definition of shareholder value, encompassing the success, flourishing, safety, preference satisfaction, and moral values of the shareholders, not…
By Stuart_ArmstrongAlignment Forum

Read the full story
Originally published by Alignment Forum on Oct 21, 2026.
Comments
Sign in to join the discussion.