S01¶


Research	R0043 — Sycophancy Vocabulary
Run	2026-04-01
Query	Q002
Search	S01

WebSearch — Regulatory standards addressing sycophancy under domain-specific names

Summary¶

Field	Value
Source/Database	WebSearch (4 queries)
Query terms	(1) EU AI Act automation bias human oversight sycophancy; (2) NIST AI 600-1 confabulation sycophancy; (3) FDA AI clinical decision support validation accuracy; (4) SR-11/7 model risk management AI
Filters	None
Results returned	40
Results selected	6
Results rejected	34

Selected Results¶

Result	Title	URL	Rationale
S01-R01	Article 14: Human Oversight - EU AI Act	https://artificialintelligenceact.eu/article/14/	Primary text of EU AI Act automation bias provision
S01-R02	NIST AI 600-1 Generative AI Profile	https://nvlpubs.nist.gov/nistpubs/ai/NIST.AI.600-1.pdf	NIST generative AI risk categories
S01-R03	SR 11-7 Model Risk Management	https://www.federalreserve.gov/supervisionreg/srletters/sr1107.htm	Federal Reserve model validation requirements
S01-R04	FDA AI/ML Guidance 2025	https://www.fda.gov/media/184856/download	FDA AI-enabled device requirements
S01-R05	IEEE 3119 AI Procurement	https://standards.ieee.org/beyond-standards/ai-procurement-guidelines-regulatory-compliance/	IEEE procurement standard for AI
S01-R06	Georgetown Tech Brief: AI Sycophancy & OpenAI	https://www.law.georgetown.edu/tech-institute/research-insights/insights/tech-brief-ai-sycophancy-openai-2/	Regulatory gap analysis for sycophancy

Rejected Results¶

Result	Title	URL	Rationale
S01-R07	Various remaining results	Various	General regulatory overviews, product marketing, news coverage, or duplicates of selected results

Notes¶

Searches confirm that no regulatory framework directly names "sycophancy" as a risk. The EU AI Act is closest with its explicit mention of "automation bias." NIST addresses "confabulation" but not agreement-seeking behavior. SR 11-7 addresses governance mechanisms (effective challenge) rather than model behaviors.