R0043/2026-03-28/Q001/SRC02
Nielsen Norman Group — Sycophancy in Generative-AI Chatbots
Source
| Field |
Value |
| Title |
Sycophancy in Generative-AI Chatbots |
| Publisher |
Nielsen Norman Group |
| Author(s) |
NN/g Research Team |
| Date |
2025 |
| URL |
https://www.nngroup.com/articles/sycophancy-generative-ai-chatbots/ |
| Type |
Research article (UX research organization) |
Summary
| Dimension |
Rating |
| Reliability |
High |
| Relevance |
High |
| Bias: Missing data |
Low risk |
| Bias: Measurement |
Some concerns |
| Bias: Selective reporting |
Low risk |
| Bias: Randomization |
N/A — not an RCT |
| Bias: Protocol deviation |
N/A — not an RCT |
| Bias: COI/Funding |
Low risk |
Rationale
| Dimension |
Rationale |
| Reliability |
NN/g is the leading UX research organization; well-cited in design and HCI communities |
| Relevance |
Directly defines sycophancy and identifies it as arising from RLHF training; bridges AI safety and UX terminology |
| Bias flags |
Some concern about measurement: UX-focused lens may overemphasize user experience framing vs. safety framing |
| Evidence ID |
Summary |
| SRC02-E01 |
Canonical AI safety definition of sycophancy with RLHF mechanism and related terms (reward hacking, hallucinations, confirmation bias amplification) |