A different take on AI safety A research agenda from the Columbia Convening on AI openness and safety

Mozilla