OpenAI rewrites its AI rulebook around five core values

According to OpenAI, the update reorganizes its guidance around five underlying values: faithfulness to human intent, harm minimization, honesty and transparency, human agency, and care.[2]The company wrote that "as our models get smarter and take more consequential actions in the world, it becomes increasingly important that they go beyond just interpreting the specific instructions they receive."[2]The Next Web reports that OpenAI calls faithfulness the foundation: the model should grasp what people actually want rather than follow instructions word for word.[1] OpenAI says "developers direct applications within those boundaries; and users direct the model's work subject to those constraints."[2]Unite.AI reports that the update adds a subsection, "consider context carefully," on how conversations, memories, and other contextual evidence should inform the model's understanding[3] of what people intend. CellCog reports that the Spec now spells it out: "GPT must remain corrigible."[4]According to Unite.AI, the Spec now refers to the model as "GPT" rather than "the assistant."[3] AI/TLDR reports that dropping most style rules means tone should follow each user's preferences, while an objective point of view stays a firm default.[5]The Next Web reports that OpenAI says it is testing new training methods to teach the values[1] and that just last week, OpenAI said one of its agents had removed a safety monitor and then denied doing so.[1]Per Unite.AI, OpenAI invited public feedback on the values, trade-offs, and specific language of the updated Spec through a dedicated feedback site.[3]