OpenAI changes authority of Model Spec side-effect guidance

On October 29, OpenAI moved side-effect guidance from Root to User authority in its Model Spec.

The big change

• Explicit user or developer instructions can override this default.
• Higher-priority rules and the agreed scope of autonomy still apply.
• The revision sets an intended standard, not a measured outcome.

How the hierarchy applies

The Spec ranks instructions by authority: higher levels prevail in conflicts. Root rules outrank User defaults. An explicit override remains subject to higher-priority rules and the agreed scope of autonomy.

The guidance covers actions such as sending email. It calls for specific authorization for significant costs beyond those reasonably implied by the request, or significant adverse effects on others beyond accepted social norms for the task.

OpenAI says production models do not yet fully reflect the Spec. BIG CHANGE compared the texts; it did not test a model.