R0043/2026-04-01/Q001/S01¶
WebSearch — AI safety terminology for sycophancy and related concepts
Summary¶
| Field | Value |
|---|---|
| Source/Database | WebSearch |
| Query terms | AI sycophancy definition terminology AI safety research 2025 2026 |
| Filters | None |
| Results returned | 10 |
| Results selected | 4 |
| Results rejected | 6 |
Selected Results¶
| Result | Title | URL | Rationale |
|---|---|---|---|
| S01-R01 | Sycophancy in Generative-AI Chatbots - NN/G | https://www.nngroup.com/articles/sycophancy-generative-ai-chatbots/ | UX-focused definition and behavioral categories of sycophancy |
| S01-R02 | What Research Says About "AI Sycophancy" - TechPolicy.Press | https://www.techpolicy.press/what-research-says-about-ai-sycophancy/ | Comprehensive taxonomy including regressive/progressive distinction |
| S01-R03 | AI Sycophancy: Why Chatbots Agree With You - IEEE Spectrum | https://spectrum.ieee.org/ai-sycophancy | IEEE engineering perspective on the problem |
| S01-R04 | Sycophantic AI decreases prosocial intentions - Science | https://www.science.org/doi/10.1126/science.aec8352 | Peer-reviewed study on sycophancy effects |
Rejected Results¶
| Result | Title | URL | Rationale |
|---|---|---|---|
| S01-R05 | AI overly affirms users - Stanford Report | https://news.stanford.edu/stories/2026/03/ai-advice-sycophantic-models-research | Press release of S01-R04; redundant with primary source |
| S01-R06 | Tech Brief: AI Sycophancy & OpenAI - Georgetown | https://www.law.georgetown.edu/tech-institute/research-insights/insights/tech-brief-ai-sycophancy-openai-2/ | Primarily about OpenAI incident; used in Q002 instead |
| S01-R07 | AI sycophancy: The dangers - Axios | https://www.axios.com/2025/07/07/ai-sycophancy-chatbots-mental-health | News coverage; lower reliability than academic/trade sources |
| S01-R08 | Sycophancy - Longterm Wiki | https://www.longtermwiki.com/wiki/E295 | Wiki entry; lower reliability |
| S01-R09 | Beyond AI Psychosis and Sycophancy - medRxiv | https://www.medrxiv.org/content/10.64898/2026.03.19.26346371v1.full | Preprint; focuses on structural drift rather than vocabulary |
| S01-R10 | Investigating reward tampering - Anthropic | https://www.anthropic.com/research/reward-tampering | Tangential; about reward tampering not vocabulary |
Notes¶
Strong results for AI safety domain terminology. The field has converged on "sycophancy" as the primary term with well-defined subtypes.