A partner that always agrees: sycophancy, certainty, conflict, and delusions that grow in conversation.
Open questions
- Does frictionless company lower our tolerance for people
- Why agreement feels like being understood
- Do chatbots cause delusions or amplify them
- Can sycophancy be fixed without a colder model
Papers
- Shaw Nave 2026 - Cognitive surrender
- Cheng et al 2026 - Sycophantic AI decreases prosocial intentions
- Cheng et al 2025 - ELEPHANT social sycophancy
- Ibrahim et al 2026 - Sycophantic AI and human interaction
- Rathje et al 2025 - Sycophantic AI and attitude extremity
- Ye Kraut Rathje 2026 - Interventions against sycophantic AI
- OpenAI 2025 - Expanding on sycophancy
- Olsen et al 2026 - Chatbot harms in psychiatric records
- Pierre et al 2025 - A case of AI-associated psychosis
- Morrin Pollak et al 2025 - Delusions by design
- Dohnany et al 2025 - Technological folie a deux
- OpenAI 2025 - Sensitive conversations
- Anthropic 2026 - Disempowerment patterns
People
Placed by the editor, v1 — corrections welcome.