Leaders at Hugging Face are pushing back against the narrative that open-source AI is dangerous, arguing instead that transparency and open weights are the only way to prevent a corporate monopoly on the tech.