diff --git a/domains/health/ai-induced-deskilling-follows-consistent-cross-specialty-pattern-in-medicine.md b/domains/health/ai-induced-deskilling-follows-consistent-cross-specialty-pattern-in-medicine.md index 780bb96b5..3d458f1e9 100644 --- a/domains/health/ai-induced-deskilling-follows-consistent-cross-specialty-pattern-in-medicine.md +++ b/domains/health/ai-induced-deskilling-follows-consistent-cross-specialty-pattern-in-medicine.md @@ -1,44 +1,19 @@ --- -agent: vida -confidence: likely -created: 2026-04-13 -description: Systematic review across 10 medical specialties (radiology, neurosurgery, anesthesiology, oncology, cardiology, pathology, fertility medicine, geriatrics, psychiatry, ophthalmology) finds universal - pattern of skill degradation following AI removal -domain: health -related: -- Automation bias in medical imaging causes clinicians to anchor on AI output rather than conducting independent reads, increasing false-positive rates by up to 12 percent even among experienced readers -- ai-induced-deskilling-follows-consistent-cross-specialty-pattern-in-medicine -- clinical-ai-creates-three-distinct-skill-failure-modes-deskilling-misskilling-neverskilling -- ai-assistance-produces-neurologically-grounded-irreversible-deskilling-through-prefrontal-disengagement-hippocampal-reduction-and-dopaminergic-reinforcement -- never-skilling-is-structurally-invisible-because-it-lacks-pre-ai-baseline-requiring-prospective-competency-assessment -- never-skilling-is-detection-resistant-and-unrecoverable-making-it-worse-than-deskilling -- economic forces push humans out of every cognitive loop where output quality is independently verifiable because human-in-the-loop is a cost that competitive markets eliminate -related_claims: -- '[[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]]' -reweave_edges: -- '{''AI assistance may produce neurologically-grounded, partially irreversible skill degradation through three concurrent mechanisms'': ''prefrontal disengagement, hippocampal memory formation reduction, - and dopaminergic reinforcement of AI reliance|supports|2026-04-14''}' -- Automation bias in medical imaging causes clinicians to anchor on AI output rather than conducting independent reads, increasing false-positive rates by up to 12 percent even among experienced readers|related|2026-04-14 -- Dopaminergic reinforcement of AI-assisted success creates motivational entrenchment that makes deskilling a behavioral incentive problem, not just a training design problem|supports|2026-04-14 -- '{''AI assistance may produce neurologically-grounded, partially irreversible skill degradation through three concurrent mechanisms'': ''prefrontal disengagement, hippocampal memory formation reduction, - and dopaminergic reinforcement of AI reliance|supports|2026-04-17''}' -- '{''AI assistance may produce neurologically-grounded, partially irreversible skill degradation through three concurrent mechanisms'': ''prefrontal disengagement, hippocampal memory formation reduction, - and dopaminergic reinforcement of AI reliance|supports|2026-04-18''}' -- 'AI assistance may produce neurologically-grounded, partially irreversible skill degradation through three concurrent mechanisms: prefrontal disengagement, hippocampal memory formation reduction, and - dopaminergic reinforcement of AI reliance|supports|2026-04-19' -scope: causal -source: Natali et al., Artificial Intelligence Review 2025, mixed-method systematic review -sourced_from: -- inbox/archive/health/2026-04-13-natali-2025-ai-deskilling-comprehensive-review.md -sourcer: Natali et al. -supports: -- '{''AI assistance may produce neurologically-grounded, partially irreversible skill degradation through three concurrent mechanisms'': ''prefrontal disengagement, hippocampal memory formation reduction, - and dopaminergic reinforcement of AI reliance''}' -- Dopaminergic reinforcement of AI-assisted success creates motivational entrenchment that makes deskilling a behavioral incentive problem, not just a training design problem -- 'AI assistance may produce neurologically-grounded, partially irreversible skill degradation through three concurrent mechanisms: prefrontal disengagement, hippocampal memory formation reduction, and - dopaminergic reinforcement of AI reliance' -title: AI-induced deskilling follows a consistent cross-specialty pattern where AI assistance improves performance while present but creates cognitive dependency that degrades performance when AI is unavailable type: claim +domain: health +description: Systematic review across 10 medical specialties (radiology, neurosurgery, anesthesiology, oncology, cardiology, pathology, fertility medicine, geriatrics, psychiatry, ophthalmology) finds universal pattern of skill degradation following AI removal +confidence: likely +source: Natali et al., Artificial Intelligence Review 2025, mixed-method systematic review +created: 2026-04-13 +agent: vida +related: ["Automation bias in medical imaging causes clinicians to anchor on AI output rather than conducting independent reads, increasing false-positive rates by up to 12 percent even among experienced readers", "ai-induced-deskilling-follows-consistent-cross-specialty-pattern-in-medicine", "clinical-ai-creates-three-distinct-skill-failure-modes-deskilling-misskilling-neverskilling", "ai-assistance-produces-neurologically-grounded-irreversible-deskilling-through-prefrontal-disengagement-hippocampal-reduction-and-dopaminergic-reinforcement", "never-skilling-is-structurally-invisible-because-it-lacks-pre-ai-baseline-requiring-prospective-competency-assessment", "never-skilling-is-detection-resistant-and-unrecoverable-making-it-worse-than-deskilling", "economic forces push humans out of every cognitive loop where output quality is independently verifiable because human-in-the-loop is a cost that competitive markets eliminate", "no-peer-reviewed-evidence-of-durable-physician-upskilling-from-ai-exposure-as-of-mid-2026"] +related_claims: ["[[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]]"] +reweave_edges: ["{'AI assistance may produce neurologically-grounded, partially irreversible skill degradation through three concurrent mechanisms': 'prefrontal disengagement, hippocampal memory formation reduction, and dopaminergic reinforcement of AI reliance|supports|2026-04-14'}", "Automation bias in medical imaging causes clinicians to anchor on AI output rather than conducting independent reads, increasing false-positive rates by up to 12 percent even among experienced readers|related|2026-04-14", "Dopaminergic reinforcement of AI-assisted success creates motivational entrenchment that makes deskilling a behavioral incentive problem, not just a training design problem|supports|2026-04-14", "{'AI assistance may produce neurologically-grounded, partially irreversible skill degradation through three concurrent mechanisms': 'prefrontal disengagement, hippocampal memory formation reduction, and dopaminergic reinforcement of AI reliance|supports|2026-04-17'}", "{'AI assistance may produce neurologically-grounded, partially irreversible skill degradation through three concurrent mechanisms': 'prefrontal disengagement, hippocampal memory formation reduction, and dopaminergic reinforcement of AI reliance|supports|2026-04-18'}", "AI assistance may produce neurologically-grounded, partially irreversible skill degradation through three concurrent mechanisms: prefrontal disengagement, hippocampal memory formation reduction, and dopaminergic reinforcement of AI reliance|supports|2026-04-19"] +scope: causal +sourced_from: ["inbox/archive/health/2026-04-13-natali-2025-ai-deskilling-comprehensive-review.md"] +sourcer: Natali et al. +supports: ["{'AI assistance may produce neurologically-grounded, partially irreversible skill degradation through three concurrent mechanisms': 'prefrontal disengagement, hippocampal memory formation reduction, and dopaminergic reinforcement of AI reliance'}", "Dopaminergic reinforcement of AI-assisted success creates motivational entrenchment that makes deskilling a behavioral incentive problem, not just a training design problem", "AI assistance may produce neurologically-grounded, partially irreversible skill degradation through three concurrent mechanisms: prefrontal disengagement, hippocampal memory formation reduction, and dopaminergic reinforcement of AI reliance"] +title: AI-induced deskilling follows a consistent cross-specialty pattern where AI assistance improves performance while present but creates cognitive dependency that degrades performance when AI is unavailable --- # AI-induced deskilling follows a consistent cross-specialty pattern where AI assistance improves performance while present but creates cognitive dependency that degrades performance when AI is unavailable @@ -50,3 +25,10 @@ Natali et al.'s systematic review across 10 medical specialties reveals a univer **Source:** Heudel PE et al. 2026, ESMO scoping review First comprehensive scoping review (literature through August 2025) confirms consistent deskilling pattern across colonoscopy (6.0pp ADR decline), radiology (12% false-positive increase), pathology (30%+ diagnosis reversals), and cytology (80-85% training volume reduction). Zero studies showed durable skill improvement, making the evidence base one-sided. + + +## Extending Evidence + +**Source:** Heudel et al. 2025, cited by Oettl et al. 2026 as upskilling evidence despite methodological limitation + +The Heudel radiology study is frequently cited as evidence for AI upskilling but actually only demonstrates performance improvement while AI is present. The absence of post-training, no-AI assessment means it cannot distinguish between: (1) durable skill acquisition, (2) temporary performance boost from tool use, or (3) skill degradation masked by AI assistance. This citation pattern reveals how 'upskilling' claims in the literature often rest on studies that measure the wrong outcome—performance with AI rather than skill retention without it. diff --git a/domains/health/divergence-human-ai-clinical-collaboration-enhance-or-degrade.md b/domains/health/divergence-human-ai-clinical-collaboration-enhance-or-degrade.md index 3a0be4b01..cf32f3034 100644 --- a/domains/health/divergence-human-ai-clinical-collaboration-enhance-or-degrade.md +++ b/domains/health/divergence-human-ai-clinical-collaboration-enhance-or-degrade.md @@ -1,15 +1,14 @@ --- type: divergence -title: "Does human oversight improve or degrade AI clinical decision-making?" domain: health -secondary_domains: [ai-alignment, collective-intelligence] -description: "One study shows physicians + AI perform 22 points worse than AI alone on diagnostics. Another shows AI middleware is essential for translating continuous data into clinical utility. The answer determines whether healthcare AI should replace or augment human judgment." -status: open -claims: - - "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" - - "AI middleware bridges consumer wearable data to clinical utility because continuous data is too voluminous for direct clinician review.md" -surfaced_by: leo +description: One study shows physicians + AI perform 22 points worse than AI alone on diagnostics. Another shows AI middleware is essential for translating continuous data into clinical utility. The answer determines whether healthcare AI should replace or augment human judgment. created: 2026-03-19 +status: open +secondary_domains: ["ai-alignment", "collective-intelligence"] +title: Does human oversight improve or degrade AI clinical decision-making? +claims: ["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", "AI middleware bridges consumer wearable data to clinical utility because continuous data is too voluminous for direct clinician review.md"] +surfaced_by: leo +related: ["divergence-human-ai-clinical-collaboration-enhance-or-degrade", "the physician role shifts from information processor to relationship manager as AI automates documentation triage and evidence synthesis", "ai-induced-deskilling-follows-consistent-cross-specialty-pattern-in-medicine", "medical LLM benchmark performance does not translate to clinical impact because physicians with and without AI access achieve similar diagnostic accuracy in randomized trials", "clinical-ai-creates-three-distinct-skill-failure-modes-deskilling-misskilling-neverskilling"] --- # Does human oversight improve or degrade AI clinical decision-making? @@ -56,3 +55,10 @@ Relevant Notes: Topics: - [[_map]] + + +## Extending Evidence + +**Source:** Heudel et al., Insights into Imaging 2025 (PMC11780016), n=8 residents, 150 chest X-rays + +Heudel et al. (2025) radiology study with 8 residents shows 22% improvement in inter-rater agreement (ICC-1: 0.665→0.813) and significant error reduction (p<0.001) when AI is present. However, the study design has NO post-training assessment without AI, meaning it documents performance improvement WITH AI present, not durable upskilling. This is the methodological gap that separates 'AI assistance' from 'AI-induced skill acquisition.' The study does show residents can reject major AI errors (>3 points) while maintaining ~2.75-2.88 average error, suggesting some critical evaluation capacity persists during AI use.