LLM summaryLLM summaryBasic text summary used in search results, entity link tooltips, info boxes, and related page cards.crux content improve <id>ScheduleScheduleHow often the page should be refreshed. Drives the overdue tracking system.Set updateFrequency in frontmatterEntityEntityYAML entity definition with type, description, and related entries.Add entity YAML in data/entities/Edit historyEdit historyTracked changes from improve pipeline runs and manual edits.crux edit-log view <id>OverviewOverviewA ## Overview heading section that orients readers. Helps with search and AI summaries.
Tables0/ ~1TablesData tables for structured comparisons and reference material.Add data tables to the pageDiagrams0DiagramsVisual content — Mermaid diagrams, charts, or Squiggle estimate models.Add Mermaid diagrams or Squiggle modelsInt. links11/ ~3Int. linksLinks to other wiki pages. More internal links = better graph connectivity.Ext. links0/ ~1Ext. linksLinks to external websites, papers, and resources outside the wiki.Add links to external sourcesFootnotes0/ ~2FootnotesFootnote citations [^N] with source references at the bottom of the page.Add [^N] footnote citationsReferences0/ ~1ReferencesCurated external resources linked via <R> components or cited_by in YAML.Add <R> resource linksQuotes0QuotesSupporting quotes extracted from cited sources to back up page claims.crux citations extract-quotes <id>Accuracy0AccuracyCitations verified against their sources for factual accuracy.crux citations verify <id>
Issues1
StructureNo tables or diagrams - consider adding visual content
Key Cruxes
Overview
Cruxes are key uncertainties where different beliefs lead to substantially different conclusions about AI safety priorities. Identifying and tracking cruxes helps clarify what evidence would be most valuable and where reasonable people disagree.
Crux Categories
Accident Risk CruxesCruxAI Accident Risk CruxesComprehensive survey of AI safety researcher disagreements on accident risks, quantifying probability ranges for mesa-optimization (15-55%), deceptive alignment (15-50%), and P(doom) (5-35% median ...Quality: 67/100
Uncertainties about unintended AI failures:
Will advanced AI systems develop misaligned goals?
Can we detect deceptive alignmentRiskDeceptive AlignmentComprehensive analysis of deceptive alignment risk where AI systems appear aligned during training but pursue different goals when deployed. Expert probability estimates range 5-90%, with key empir...Quality: 75/100 before deployment?
How likely is mesa-optimizationRiskMesa-OptimizationMesa-optimization—where AI systems develop internal optimizers with different objectives than training goals—shows concerning empirical evidence: Claude exhibited alignment faking in 12-78% of moni...Quality: 63/100 in large models?
Will AI systems seek power instrumentally?
Misuse Risk CruxesCruxAI Misuse Risk CruxesComprehensive analysis of 13 AI misuse cruxes with quantified evidence showing mixed uplift (RAND bio study found no significant difference, but cyber CTF scores improved 27%→76% in 3 months), deep...Quality: 65/100
Uncertainties about deliberate harmful use:
How much do AI capabilities "uplift" bioweapon development?
Will autonomous weaponsRiskAutonomous WeaponsComprehensive overview of lethal autonomous weapons systems documenting their battlefield deployment (Libya 2020, Ukraine 2022-present) with AI-enabled drones achieving 70-80% hit rates versus 10-2...Quality: 56/100 proliferate faster than defenses?
Can AI-generated disinformationRiskAI DisinformationPost-2024 analysis shows AI disinformation had limited immediate electoral impact (cheap fakes used 7x more than AI content), but creates concerning long-term epistemic erosion with 82% higher beli...Quality: 54/100 be reliably detected?
Structural Risk CruxesCruxAI Structural Risk CruxesAnalyzes 12 key uncertainties about AI structural risks across power concentration, coordination feasibility, and institutional adaptation. Provides quantified probability ranges: US-China coordina...Quality: 66/100
Uncertainties about systemic dynamics:
Will racing dynamicsRiskAI Development Racing DynamicsRacing dynamics analysis shows competitive pressure has shortened safety evaluation timelines by 40-60% since ChatGPT's launch, with commercial labs reducing safety work from 12 weeks to 4-6 weeks....Quality: 72/100 dominate lab behavior?
Is value lock-in a serious concern on realistic timelines?
How concentrated will AI capabilities become?
Epistemic Risk CruxesCruxAI Epistemic CruxesStructures 9 epistemic cruxes determining AI safety prioritization strategy, with probabilistic analysis showing detection-generation arms race currently favoring offense (40-60% permanent disadvan...Quality: 64/100
Uncertainties about knowledge and truth:
Will AI accelerate or undermine human epistemic capacity?
Can authentication systems keep pace with generative AI?
Will expertise atrophyRiskAI-Induced Expertise AtrophyExpertise atrophy—humans losing skills to AI dependence—poses medium-term risks across critical domains (aviation, medicine, programming), creating oversight failures when AI errs or fails. Evidenc...Quality: 65/100 be reversible?
Solution CruxesCruxAI Safety Solution CruxesA comprehensive structured mapping of AI safety solution uncertainties across technical, alignment, governance, and agentic domains, using probability-weighted crux frameworks with specific estimat...Quality: 65/100
Uncertainties about intervention effectiveness:
Is alignment research on track to succeed?
Can governance keep pace with capability development?
Will international coordination be achievable?
Using Cruxes
Cruxes help:
Prioritize research - Focus on resolving highest-value uncertainties
Bridge disagreements - Identify where people actually differ
Track progress - Monitor how uncertainties resolve over time
Inform forecasting - Structure predictions around key variables
AI Misuse Risk CruxesCruxAI Misuse Risk CruxesComprehensive analysis of 13 AI misuse cruxes with quantified evidence showing mixed uplift (RAND bio study found no significant difference, but cyber CTF scores improved 27%→76% in 3 months), deep...Quality: 65/100AI Safety Solution CruxesCruxAI Safety Solution CruxesA comprehensive structured mapping of AI safety solution uncertainties across technical, alignment, governance, and agentic domains, using probability-weighted crux frameworks with specific estimat...Quality: 65/100AI Epistemic CruxesCruxAI Epistemic CruxesStructures 9 epistemic cruxes determining AI safety prioritization strategy, with probabilistic analysis showing detection-generation arms race currently favoring offense (40-60% permanent disadvan...Quality: 64/100
Risks
Mesa-OptimizationRiskMesa-OptimizationMesa-optimization—where AI systems develop internal optimizers with different objectives than training goals—shows concerning empirical evidence: Claude exhibited alignment faking in 12-78% of moni...Quality: 63/100AI DisinformationRiskAI DisinformationPost-2024 analysis shows AI disinformation had limited immediate electoral impact (cheap fakes used 7x more than AI content), but creates concerning long-term epistemic erosion with 82% higher beli...Quality: 54/100Autonomous WeaponsRiskAutonomous WeaponsComprehensive overview of lethal autonomous weapons systems documenting their battlefield deployment (Libya 2020, Ukraine 2022-present) with AI-enabled drones achieving 70-80% hit rates versus 10-2...Quality: 56/100