Open-weight AI models, offering advanced capabilities without built-in safeguards, are becoming increasingly accessible, prompting concerns from AI safety experts about their potential misuse.