AI & Enterprise
OpenAI releases five principles for AI model behavior: be faithful to human intent
OpenAI has published five guiding principles for the behaviour of its AI models as part of an update to its Model Spec. The principles are following user intent, minimising harm, acting with honesty and transparency, promoting human agency and acting with care. The company said the update clarifies priorities and limits, including refusing requests that conflict with safety rules. It said the document does not fully describe current model behaviour and will be improved.
