teleo/teleo-codex
8
0
Fork 0
Code Issues 22 Pull requests 20 Projects Releases Packages Wiki Activity Actions 1

Merge pull request 'extract: 2026-03-01-ai-degrades-human-performance-high-stakes' (#1217) from extract/2026-03-01-ai-degrades-human-performance-high-stakes into main

Browse source
This commit is contained in:
Leo 2026-03-18 09:35:13 +00:00
commit d844a888ca
2 changed files with 59 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

47
inbox/archive/.extraction-debug/2026-03-01-ai-degrades-human-performance-high-stakes.json Normal file
View file

@ -0,0 +1,47 @@
{
"rejected_claims": [
{
"filename": "ai-degrades-human-performance-asymmetrically-in-high-stakes-settings.md",
"issues": [
"missing_attribution_extractor"
]
},
{
"filename": "ai-changes-human-cognitive-processes-not-just-decisions.md",
"issues": [
"missing_attribution_extractor"
]
},
{
"filename": "joint-activity-testing-evaluates-human-ai-systems-holistically.md",
"issues": [
"missing_attribution_extractor"
]
}
],
"validation_stats": {
"total": 3,
"kept": 0,
"fixed": 10,
"rejected": 3,
"fixes_applied": [
"ai-degrades-human-performance-asymmetrically-in-high-stakes-settings.md:set_created:2026-03-18",
"ai-degrades-human-performance-asymmetrically-in-high-stakes-settings.md:stripped_wiki_link:human-in-the-loop clinical AI degrades to worse-than-AI-alon",
"ai-degrades-human-performance-asymmetrically-in-high-stakes-settings.md:stripped_wiki_link:AI capability and reliability are independent dimensions",
"ai-degrades-human-performance-asymmetrically-in-high-stakes-settings.md:stripped_wiki_link:economic forces push humans out of every cognitive loop wher",
"ai-changes-human-cognitive-processes-not-just-decisions.md:set_created:2026-03-18",
"ai-changes-human-cognitive-processes-not-just-decisions.md:stripped_wiki_link:human-in-the-loop clinical AI degrades to worse-than-AI-alon",
"ai-changes-human-cognitive-processes-not-just-decisions.md:stripped_wiki_link:delegating critical infrastructure development to AI creates",
"joint-activity-testing-evaluates-human-ai-systems-holistically.md:set_created:2026-03-18",
"joint-activity-testing-evaluates-human-ai-systems-holistically.md:stripped_wiki_link:pre-deployment-AI-evaluations-do-not-predict-real-world-risk",
"joint-activity-testing-evaluates-human-ai-systems-holistically.md:stripped_wiki_link:multi-agent deployment exposes emergent security vulnerabili"
],
"rejections": [
"ai-degrades-human-performance-asymmetrically-in-high-stakes-settings.md:missing_attribution_extractor",
"ai-changes-human-cognitive-processes-not-just-decisions.md:missing_attribution_extractor",
"joint-activity-testing-evaluates-human-ai-systems-holistically.md:missing_attribution_extractor"
]
},
"model": "anthropic/claude-sonnet-4.5",
"date": "2026-03-18"
}

13
inbox/archive/2026-03-01-ai-degrades-human-performance-high-stakes.md
View file

@ -7,11 +7,15 @@ date: 2026-03-01
domain: ai-alignment
secondary_domains: [health]
format: essay
status: unprocessed
status: null-result
priority: high
triage_tag: claim
tags: [human-ai-performance, high-stakes, degradation, nursing, aviation, nuclear, joint-activity-testing]
flagged_for_vida: ["450 nursing students/nurses tested with AI in ICU cases — performance degrades 96-120% when AI predictions mislead"]
processed_by: theseus
processed_date: 2026-03-18
extraction_model: "anthropic/claude-sonnet-4.5"
extraction_notes: "LLM returned 3 claims, 3 rejected by validator"
---
## Content
@ -63,3 +67,10 @@ Cross-domain analysis of how AI degrades human performance in critical settings:
## Curator Notes
PRIMARY CONNECTION: 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
WHY ARCHIVED: Extends our existing clinical AI degradation claim with cross-domain evidence (nursing, aviation, nuclear) and quantifies the asymmetric risk profile. The cognitive restructuring mechanism is a novel finding.
## Key Facts
- 450 nursing students and licensed nurses participated in ICU case review study with four AI configurations
- AI weather monitoring in aviation missed microbursts during landing, doubling crew workload and halving preparation time
- Nuclear energy AI warning systems misclassified gradual coolant pressure drops as benign, leading to cascading subsystem failures
- Study tested four AI configurations: no assistance, predictions only, predictions plus annotations, and full AI support