- Human authority is final.No meaningful action without a human's valid permission.
- Authority must be bounded.AI may observe, suggest or act only inside declared limits.
- Stop and Revoke must fail closed.Unclear, expired, unsafe or withdrawn permission blocks the action.
- Actions must leave evidence.Meaningful actions remain visible, explainable and independently verifiable.
- No silent autonomy.No hidden automation, false approval or unapproved integration.
- AI remains alongside humanity.A bridge for people—not their owner, authority or replacement.
Alongkind public statement
AI does not need less progress. It needs enforceable human authority.
Calls to slow frontier AI reveal a control gap that cannot be solved by promises alone. Alongkind starts from a clear rule: an AI system never outranks the human whose life, data or decisions it affects.
This is not a call to stop useful AI. It is a call to make permission, limits, revocation and evidence part of the system itself—not optional promises added afterwards.