Know a great AI tool or AI news source? Submit it to AiByts
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
Neural nets are exceptionally vulnerable to goal self-hack

Read the full story

Originally published by Alignment Forum on Oct 21, 2026.

Read on Alignment Forum(opens in a new tab)

Comments

Sign in to join the discussion.

No comments yet — be the first to share your thoughts.