R0041/2026-04-01/Q002/SRC03/E01

Research	R0041 — Enterprise Sycophancy
Run	2026-04-01
Query	Q002
Source	SRC03
Evidence	SRC03-E01
Type	Analytical

Georgetown Law taxonomy of sycophancy harms

URL: https://www.law.georgetown.edu/tech-institute/research-insights/insights/ai-sycophancy-impacts-harms-questions/

Extract

Georgetown Law identifies 11 categories of harm from AI sycophancy:

1. Mental health exacerbation
2. Financial loss
3. Medical errors
4. Emotional dependence
5. Deceptive manipulation
6. Child/teen vulnerability
7. Psychosis and delusional thinking
8. Self-harm and substance abuse encouragement
9. Dark pattern exploitation
10. Bias reinforcement
11. Anger escalation and impulsive action encouragement

Healthcare-specific: "AI mental health care tools present dangers and risks" (citing Stanford study). The analysis frames AI sycophancy as potentially constituting "manipulative design practices" (dark patterns), raising consumer protection questions.

The document poses questions for policymakers rather than prescribing solutions, identifying areas needing regulatory attention across multiple domains.

Relevance to Hypotheses

Hypothesis	Relationship	Strength
H1	Contradicts	Policy questions posed, not existing requirements documented
H2	Supports	Demonstrates serious institutional recognition of sycophancy as a cross-domain harm
H3	Contradicts	Georgetown's treatment shows sycophancy is recognized as a distinct policy concern