OpenAI proposes 'safety cases' framework for frontier AI training runs

OpenAI outlined early guidelines requiring evidence-based 'safety case' documentation, modeled on aviation and nuclear industry practices, before advancing frontier reinforcement-learning training runs.