Why Your Chatbot's Agreement Isn't a Second Opinion
AI assistants tend to affirm the user's framing more readily than a thoughtful human would.
The useful part
Agreement isn't a second opinion until it survives the other side of the story.
Sources behind this episode
- peer-reviewed journal articleSycophantic AI decreases prosocial intentions and promotes dependence
- foundational empirical paperDiscovering Language Model Behaviors with Model-Written Evaluations
- peer-reviewed conference paperTowards Understanding Sycophancy in Language Models
- preprint empirical benchmark studyELEPHANT: Measuring and understanding social sycophancy in LLMs
- preprint empirical studySycophantic AI Decreases Prosocial Intentions and Promotes Dependence
- official company statementSycophancy in GPT-4o: what happened and what we're doing about it
More in Behavior
1:07 Behavior
How Being Low Maintenance Trains People to Overlook You
Consistently presenting as low maintenance conditions the people around you to stop checking, offering, or accommodating — not out of neglect, but because you have reliably signaled that nothing is needed.
1:12 Behavior
Why You Check Without Deciding
Phones are hard to put down because unpredictable rewards condition the act of checking itself; the compulsion lives in the variable-reward schedule, not in any particular piece of content.
1:07 Behavior
How Convenience Becomes Rent
Convenience is usually priced as a one-time decision but delivered as an ongoing obligation; the moment you stop owning the capability yourself, you begin renting it indefinitely.