AI chatbot safeguards fail to prevent spread of health disinformation, study reveals
A study assessed the effectiveness of safeguards in foundational large language models (LLMs) to protect against malicious instruction that could turn them into tools for spreading disinformation, or the deliberate…