domains/health/human-in-the-loop clinical AI degrades to worse-than-AI-alone because physicians both de-skill from reliance and introduce errors when overriding correct outputs.md

@@ -38,6 +38,12 @@ OpenEvidence's 1M daily consultations (30M+/month) with 44% of physicians expres
 
 The Sutter Health-OpenEvidence EHR integration creates a natural experiment in automation bias: the same tool (OpenEvidence) that was previously used as an external reference is now embedded in primary clinical workflows. Research on in-context vs. external AI shows in-workflow suggestions generate higher adherence, suggesting the integration will increase automation bias independent of model quality changes.
 
+### Additional Evidence (extend)
+*Source: [[2026-03-09-mount-sinai-multi-agent-clinical-ai-nphealthsystems]] | Added: 2026-03-23*
+
+Mount Sinai's multi-agent architecture study (npj Health Systems, March 2026) provides the first peer-reviewed clinical demonstration of an alternative architecture that avoids human-in-the-loop degradation. By distributing tasks among specialized agents rather than requiring human oversight of a single generalist model, the system maintains performance under heavy workload where single-agent systems degrade. This suggests the solution to HITL degradation is architectural (multi-agent specialization) rather than training-based (better models) or procedural (better human oversight protocols).
+
+
 
 Relevant Notes:
 - [[centaur team performance depends on role complementarity not mere human-AI combination]] -- the chess centaur model does NOT generalize to clinical medicine where physician overrides degrade AI performance

inbox/queue/.extraction-debug/2026-03-09-mount-sinai-multi-agent-clinical-ai-nphealthsystems.json

@@ -0,0 +1,35 @@
+{
+  "rejected_claims": [
+    {
+      "filename": "multi-agent-clinical-ai-reduces-computational-cost-65x-while-maintaining-performance-under-workload.md",
+      "issues": [
+        "missing_attribution_extractor"
+      ]
+    },
+    {
+      "filename": "multi-agent-clinical-ai-adoption-driven-by-efficiency-not-safety-creating-accidental-harm-reduction.md",
+      "issues": [
+        "missing_attribution_extractor"
+      ]
+    }
+  ],
+  "validation_stats": {
+    "total": 2,
+    "kept": 0,
+    "fixed": 5,
+    "rejected": 2,
+    "fixes_applied": [
+      "multi-agent-clinical-ai-reduces-computational-cost-65x-while-maintaining-performance-under-workload.md:set_created:2026-03-23",
+      "multi-agent-clinical-ai-reduces-computational-cost-65x-while-maintaining-performance-under-workload.md:stripped_wiki_link:human-in-the-loop clinical AI degrades to worse-than-AI-alon",
+      "multi-agent-clinical-ai-adoption-driven-by-efficiency-not-safety-creating-accidental-harm-reduction.md:set_created:2026-03-23",
+      "multi-agent-clinical-ai-adoption-driven-by-efficiency-not-safety-creating-accidental-harm-reduction.md:stripped_wiki_link:human-in-the-loop clinical AI degrades to worse-than-AI-alon",
+      "multi-agent-clinical-ai-adoption-driven-by-efficiency-not-safety-creating-accidental-harm-reduction.md:stripped_wiki_link:healthcare AI creates a Jevons paradox because adding capaci"
+    ],
+    "rejections": [
+      "multi-agent-clinical-ai-reduces-computational-cost-65x-while-maintaining-performance-under-workload.md:missing_attribution_extractor",
+      "multi-agent-clinical-ai-adoption-driven-by-efficiency-not-safety-creating-accidental-harm-reduction.md:missing_attribution_extractor"
+    ]
+  },
+  "model": "anthropic/claude-sonnet-4.5",
+  "date": "2026-03-23"
+}

inbox/queue/2026-03-09-mount-sinai-multi-agent-clinical-ai-nphealthsystems.md

@@ -7,9 +7,13 @@ date: 2026-03-09
 domain: health
 secondary_domains: [ai-alignment]
 format: research paper
-status: unprocessed
+status: enrichment
 priority: high
 tags: [clinical-ai-safety, multi-agent-ai, efficiency, noharm, agentic-ai, healthcare-workflow, atoms-to-bits, belief-5]
+processed_by: vida
+processed_date: 2026-03-23
+enrichments_applied: ["human-in-the-loop clinical AI degrades to worse-than-AI-alone because physicians both de-skill from reliance and introduce errors when overriding correct outputs.md"]
+extraction_model: "anthropic/claude-sonnet-4.5"
 ---
 
 ## Content
@@ -58,3 +62,11 @@ Published online March 9, 2026 in npj Health Systems. Senior author: Girish N. N
 PRIMARY CONNECTION: "human-in-the-loop clinical AI degrades to worse-than-AI-alone" — multi-agent is the architectural counter-proposal; this paper is the first commercial-grade evidence for that architecture
 WHY ARCHIVED: First peer-reviewed demonstration of multi-agent clinical AI entering healthcare deployment; the framing gap (efficiency vs. safety) is a new KB finding about how research evidence translates to market adoption
 EXTRACTION HINT: Extract two claims: (1) multi-agent architecture outperforms single-agent on efficiency AND performance in healthcare; (2) multi-agent is being adopted for efficiency reasons not safety reasons, creating a paradoxical situation where NOHARM's safety case may be implemented accidentally via cost-reduction adoption. The second claim requires care — it's an inference, should be "experimental."
+
+
+## Key Facts
+- Mount Sinai study evaluated multi-agent AI across patient information retrieval, clinical data extraction, and medication dose checking
+- Study published March 9, 2026 in npj Health Systems
+- Coverage by EurekAlert!, Medical Xpress, NewsWise, and News-Medical
+- Dr. Nathan Moore demonstrated multi-agent for end-of-life and advance care planning automation at HIMSS 2026 Global Health Conference
+- BCG published 'AI agents will transform health care in 2026' report in January 2026