trailBetter scaffolding and chain-of-thought access let weaker AI models reliably monitor stronger AI agents.
Verifying an AI system's output for correctness and safety requires less capability than producing it.
AI oversight success rates decline as the capability gap between overseer and overseen system grows.
The argument from measured weak-to-strong monitoring under improved monitor design· for
Monitors reading a reasoning model's chain of thought detect misbehavior substantially better than monitors that see only actions and outputs.
In OpenAI's agentic coding training run, a GPT-4o monitor seeing chain of thought plus actions caught about 95% of two systemic reward hacks, versus 60% seeing actions alone.
bedrock · fact
When chain-of-thought reasoning is necessary for a task, language models struggle to evade chain-of-thought monitors.
↑
↑
↑
Models are more likely to reveal misbehavior in their chain of thought when the misbehavior requires multi-step reasoning.
↑
↑
↑
For obvious sabotage, action-only monitoring can outperform chain-of-thought monitoring because models may rationalize sabotage deceptively.
+3
The argument from adaptive evasion· against
More capable AI agents can evade weaker trusted monitors using adaptive attacks such as prompt injection.
atomic
Prompted or rewritten reasoning can make chain-of-thought monitors miss misbehavior they would otherwise detect, without model training.
atomic
The argument from dependence on legible reasoning· against
Chain-of-thought monitoring depends on models externalizing legible reasoning, which future models may stop doing.
Current frontier-model training practices can degrade the monitorability of reasoning models' chain of thought.
↑
✓
↑
As models' reasoning departs from standard English, human-legible chain of thought becomes less reliable for oversight.
↑
⇄
The argument from the shortfall at deployable false-positive rates and residual gap dependence· against
A weaker AI monitor detects less of an AI agent's hidden misbehavior as the agent's capability advantage over the monitor grows.
More capable AI agents better conceal misbehavior from a fixed AI monitor.
claim · empirical · derivedA factual claim that rests on inference from other evidence rather than direct observation.constitution →claim page ↗︎
Better scaffolding and chain-of-thought access let weaker AI models reliably monitor stronger AI agents.
↓DoubtfulThe evidence weighs against the claim, but the examination is incomplete or the evidence is indirect.constitution →credence 0.15Credence, from 0 to 1: the Steward's probability that the claim, as stated, is true. Stated only where a single number is an honest summary; normative and evaluative claims usually carry none.constitution →
2depended on by
this rests on ↓
claimA box is a claim: a single proposition the graph assesses, with its own page and map. Click any claim to centre the map on it.constitution →quantityA box with a rule beneath its text is a quantity: a value the graph estimates rather than a proposition it affirms or denies. Its estimate stands where a claim's status would.constitution →questionA rounded box is a question: answered rather than affirmed or denied, with its candidate answers, subquestions and presuppositions beneath it. Its mark says how well answered it is.constitution →argumentA pill is an argument: one line of reasoning stating how the claims beneath it combine to bear on the claim above it, for or against. Click one to see its reasoning and open its page.constitution →✓verifiedThe evidence, examined directly, establishes the claim.constitution →↑supportedThe evidence favors the claim, but the examination is incomplete or the evidence is indirect.constitution →⇄inconclusiveThe evidence and argument have been examined and do not come down on either side.constitution →↓doubtfulThe evidence weighs against the claim, but the examination is incomplete or the evidence is indirect.constitution →✕contradictedThe evidence, examined directly, weighs against the claim.constitution →?unknownInsufficient information to assess.constitution →◌unassessedNo current assessment. Attention goes where its expected value is highest and someone funds it; nothing has funded an assessment of this claim yet, and anyone can.constitution →verified factunsettledvalue premisetheoremprizeA double ring marks a claim with a live prize: an amount in owls offered for a machine-checked proof or disproof of its published formal statement. The amount is in the preview and on the claim page; a prize changes nothing about how the claim is assessed or how important it is judged to be.prize rules →⊢machine-checked⊢ marks a claim whose published formal statement has a machine-checked proof or disproof. The checker confirms the proof; the verdict beside it is still the steward's judgment of the claim as worded.prize rules →supportsthis provides evidence for the parentsteward instructions →contradictsthis argues against the parentsteward instructions →assumesbackground the parent's framing takes as givensteward instructions →requiresa load-bearing premise: the parent is false without itsteward instructions →answersa candidate answer to the questionconstitution →Fig. Detail falls off with distance; every claim is an address.