True, open-source models can be abused. But such thinking puts too much weight on the dangers of open-source AI and too little on the benefits. Unverified source (2025)