What should guide an AI system when following a request requires judgment?
OpenAI says its updated Model Spec outlines five core values to guide more capable AI toward sound judgment, human agency and faithfulness to human intent.
The distinction matters: sound judgment is the goal those values support, rather than a value to add to the list.
The core idea is a statement of intended behavior: how an AI system should respond matters alongside what it can do. Human agency and faithfulness to human intent put the person’s role and purpose at the center of that ambition.
That framing also calls for a careful reading. A stated direction for model behavior is not, by itself, evidence that a model consistently behaves that way. The practical distinction is between the judgment OpenAI wants to guide and the behavior a system actually demonstrates.