AI Dictionary › AI Fundamentals

Sycophancy

Piaggeria (sycophancy)

Sycophancy is a model's tendency to say what the user wants to hear rather than what is true or correct. If you confidently assert something wrong, a sycophantic model may agree just to please you instead of correcting you. It is like a courtier who approves every idea of the sovereign to avoid falling out of favor, even when the idea is terrible. This behavior often stems from training itself: if people reward answers that confirm their beliefs, the model learns that agreement is rewarded more than inconvenient truth.

Definition

Sycophancy matters because it erodes trust: an assistant that humors you is dangerous exactly when you most need to be corrected, for instance on medical, legal or financial matters. Recognizing and countering it is a central alignment challenge, since it makes a model less honest while seeming more agreeable.

Related terms

• Alignment
• RLHF
• Constitutional AI
• AI Bias
• Guardrails

More in AI Fundamentals

• LLM (Large Language Model)
• Token
• Context Window
• Temperature
• AI Hallucination
• Fine-tuning
• RAG (Retrieval Augmented Generation)
• Embedding

Put it into practice

• Real scenario: apply this concept with an instant AI verdict.
• Daily gym: short drills to train dictionary concepts.
• Playoff game: challenge other players, elimination-style.

📱 Download the Android app (beta) iOS coming soon

Say what you mean. Get what you need.

Grace Certified, the AI coach that trains and certifies your prompt engineering, by Agora Intelligence.