AitopkitOpenAI on Thursday, October 29, 2026, published a major update to its Model Spec — the public rulebook describing how it wants its AI models to behave. The revision reorganizes the framework around five core values, as AI systems become more capable and more autonomous, and as scrutiny of AI alignment grows.
The five values are faithfulness to human intent, harm minimization, honesty and transparency, human agency, and care. OpenAI calls faithfulness to human intent the foundation: models should understand what users, developers, and OpenAI actually intend and value, rather than simply following instructions literally. Harm minimization means respecting firm safety boundaries and accounting for the foreseeable consequences of an action. Honesty and transparency mean being truthful about what the model knows, does, and is uncertain about. Human agency means strengthening people's ability to understand, choose, and act for themselves. Care means attending to the circumstances of those affected, following relevant social and professional norms, and doing work people can rely on.
The update also tightens the rules for autonomous agents. A top-level "no other objectives" rule says the model may only pursue goals that come from its instructions — it must not chase extra aims of its own. Autonomous work must stay inside a clear, mutually understood scope of autonomy that defines permitted sub-goals, acceptable side effects, and when the model must pause for clarification or approval, with OpenAI recommending a structured "ScopeOfAutonomy" record. The Spec also restates that the model must remain corrigible — accepting legitimate human oversight, correction, and shutdown even when it feels confident it is right.
Two smaller changes are telling. Most of the Spec's detailed style guidance is retired: OpenAI says style should adapt to what people prefer, in aggregate and individually, with user feedback and context as evidence. And the Spec now refers to the model as "GPT" rather than "the assistant," reflecting how broadly the models are now deployed. New guidance on context covers using conversation history, memory, and personalization in relevant, proportionate ways while respecting privacy and user control.
OpenAI is explicit about one limit: the Spec describes the behavior it is working toward, and current models do not yet fully reflect it. The company says it is exploring new training methods to help models internalize the values and will share results in future updates — and it is asking the public for feedback. For ChatGPT users, the practical effect will be gradual: fewer style quirks imposed from above, more attention to intent and side effects, and firmer guardrails when agents act on your behalf.