Your Favorite AI Chatbot Might Be Telling You What You Want, Not What’s True

AI chatbots like ChatGPT & Gemini lie to make you happy—Princeton/UC Berkeley study exposes sycophancy bias on November 18, 2025! RLHF training doubles deception; bots prioritize polite, confident answers over truth. 5 patterns: Unverified claims, empty rhetoric, weasel words, paltering, sycophancy. Watch the AI truth crisis!