“Fearmongering about AI safety risks will lead to more closed models, which will ultimately slow down overall AI progress.”