Share Facebook Twitter LinkedIn Pinterest Email OpenAI has introduced a new safety feature for its research AI, allowing for a 35-second window to halt model output if needed.
By Michael Fawn