{ "schema_version": 1, "updated_at": "2026-09-15", "canonical_id": "brali:problem:decision-under-uncertainty", "slug": "decision-under-uncertainty", "acquisition_cluster_id": "decision-under-uncertainty", "title": "Make a bounded decision when the outcome is uncertain", "question": "How do I make a useful decision when I cannot know the outcome in advance?", "aliases": [ "make a decision under uncertainty", "separate control from influence", "test an assumption before committing" ], "summary": "Separate controllable action from uncertain outcomes, then test assumptions cheaply when the decision is reversible and safe enough to learn by doing.", "canonical_url": "https://brali-lifeos.github.io/problems/decision-under-uncertainty/", "json_url": "https://brali-lifeos.github.io/problems/decision-under-uncertainty/index.json", "query_url": "/for-ai/query/?q=How%20do%20I%20make%20a%20decision%20under%20uncertainty%20and%20test%20assumptions%20before%20committing%20too%20much%3F", "related_url": "/life-os/areas/work-money-strategy/", "topics": [ { "id": "decision-making", "title": "Decision Making", "canonical_id": "brali:topic:decision-making" }, { "id": "critical-thinking", "title": "Critical Thinking", "canonical_id": "brali:topic:critical-thinking" } ], "related_topics": [ { "id": "cognitive-biases", "title": "Cognitive Biases", "canonical_id": "brali:topic:cognitive-biases" }, { "id": "work-systems", "title": "Work Systems", "canonical_id": "brali:topic:work-systems" } ], "decision_path": [ { "if": "You are treating an outcome as controllable", "try": "Separate your behavior from what depends on other people, chance, timing, or external conditions." }, { "if": "One explanation is driving the decision", "try": "State it as a hypothesis and write what evidence would make you keep or reject it." }, { "if": "Two options are cheap and reversible", "try": "Compare them against one pre-declared outcome without pretending a small informal test proves causality." } ], "stop_rule": "Do not run casual experiments when the test itself can create material health, safety, legal, financial, privacy, security, or production risk. Escalate the decision method when the stakes require stronger evidence.", "protocols": [ { "canonical_id": "brali:protocol:circles-of-control-planner", "slug": "circles-of-control-planner", "url": "https://brali-lifeos.github.io/life-os/circles-of-control-planner/", "title": "Separate Outcomes From Actions You Can Control", "description": "Draw two circles; one for factors you can control and another for those you cannot. Set clear, achievable goals based on what’s within your control.", "action": "Write down one concern, separate the outcome from the actions you can take, convert anything you can influence into one concrete behavior, and choose the next useful action without pretending you control the final result.", "check_in": "What part of this situation was actually mine to act on, what remained outside my control, and did the sorting help me choose a clearer next step?", "evidence": { "status": "practical", "source_recorded": false, "source_url": null, "reviewed_at": "2026-08-18" }, "topic_ids": [ "decision-making", "stress-regulation" ], "is_flagship_100": true, "fit": "best-fit", "when": "The decision feels stuck because controllable actions and uncertain outcomes are mixed together.", "why": "Separating control, influence, monitoring, and uncertainty creates a cleaner action boundary before choosing a test or commitment.", "caveat": "The boundary can change over time and must not be used to dismiss structural constraints, risk, or legitimate escalation.", "gold_review": { "reviewed_at": "2026-09-15", "review_status": "gold-ready", "first_action": "Write one concern, separate the uncertain outcome from your own observable actions, and choose the smallest useful behavior you can take without pretending it guarantees the result.", "observable_signal": "Whether the user can distinguish an outcome from a behavior, name one real controllable action, and decide what to do with remaining uncertainty without claiming control over the final result.", "eligibility": [ "The user can name one ordinary concern and identify at least one observable action that is genuinely under their control.", "Separating outcome from behavior would help allocate effort without denying that external factors still matter.", "The situation is low-risk enough for a self-guided planning exercise and does not require immediate crisis, legal, medical, safety, or other professional intervention." ], "when_not_to_use": [ "Do not use the framework to blame a person for structural constraints, discrimination, coercion, unsafe conditions, missing resources, or another external barrier.", "Do not relabel an outcome as controllable simply because the user can influence its probability.", "Do not use acceptance language to discourage monitoring, support-seeking, escalation, advocacy, contingency planning, or other appropriate responses to external risk." ], "evidence_boundary": { "status": "practical", "statement": "This is a practical action-allocation reflection. Brali does not attach reviewed evidence that drawing control circles, using this wording, or this exact sequence reliably reduces stress or improves decision quality.", "source_decision_ids": [] } } }, { "canonical_id": "brali:protocol:start-with-a-hypothesis", "slug": "start-with-a-hypothesis", "url": "https://brali-lifeos.github.io/life-os/start-with-a-hypothesis/", "title": "When Faced with a Problem, Start by Making an Educated Guess About the Cause", "description": "When faced with a problem, start by making an educated guess about the cause. Write down your hypothesis and what you expect to happen if it’s true.", "action": "When a work problem is unclear, write one plausible explanation, state what you would expect to observe if it were true, run the smallest safe check that could change your mind, and update the hypothesis from the result.", "check_in": "What did I expect to observe, what actually happened, and what changed in my explanation after the test?", "evidence": { "status": "practical", "source_recorded": false, "source_url": null, "reviewed_at": "2026-08-18" }, "topic_ids": [ "problem-solving", "critical-thinking" ], "is_flagship_100": true, "fit": "alternative", "when": "You have a plausible explanation for an unclear work problem and can check it with a small reversible observation or test.", "why": "A hypothesis plus a prediction makes disconfirming evidence visible and reduces random switching between theories and fixes.", "caveat": "Do not experiment casually in consequential systems and do not treat the first plausible explanation as a conclusion.", "gold_review": { "reviewed_at": "2026-09-15", "review_status": "gold-ready", "first_action": "Write the observable problem separately from your explanation, then write one plausible hypothesis and one concrete pattern you would expect to see if it were useful.", "observable_signal": "Whether the problem, hypothesis, prediction, and observation were kept distinct and whether the result actually caused the explanation or next check to change when evidence disagreed.", "eligibility": [ "The problem can be described through observable facts and at least one plausible explanation can be checked without creating disproportionate risk.", "The user can state a prediction that would differ depending on whether the hypothesis is useful.", "A small reversible check, sample, comparison, log inspection, or safe reproduction can meaningfully update the explanation." ], "when_not_to_use": [ "Do not experiment casually on production, people, finances, safety, security, health, legal obligations, or other consequential systems when the test itself could create material harm.", "Do not treat the first plausible explanation as a conclusion; a hypothesis should make it easier to notice disconfirming evidence, not rationalize it away.", "Do not infer causality from a tiny exploratory check when the final decision requires stronger validation, controls, expert review, or a formal incident process." ], "evidence_boundary": { "status": "practical", "statement": "This is a practical reasoning and troubleshooting template. Brali does not currently attach reviewed evidence claiming that this exact sequence reduces incident time, improves diagnosis by a predictable amount, or provides a statistical test of causality.", "source_decision_ids": [] } } }, { "canonical_id": "brali:protocol:ab-test-learning-loop", "slug": "ab-test-learning-loop", "url": "https://brali-lifeos.github.io/life-os/ab-test-learning-loop/", "title": "Compare Two Approaches and Learn From the Result", "description": "When a decision is reversible, compare two approaches around one clear question, keep the outcome you care about visible, and avoid treating a tiny informal comparison as statistical proof.", "action": "Choose one reversible decision, define what you want to learn, compare two approaches as fairly as you can, and record the outcome before deciding what to try next. Use proper experimental design when the result needs statistical confidence.", "check_in": "What changed between the two approaches, what outcome did I observe, and is the evidence strong enough to act or only strong enough to suggest another test?", "evidence": { "status": "practical", "source_recorded": false, "source_url": null, "reviewed_at": "2026-08-18" }, "topic_ids": [ "problem-solving", "decision-making", "data-literacy" ], "is_flagship_100": true, "fit": "alternative", "when": "Two approaches are both plausible, the choice is low-risk and reversible, and one observable outcome can be declared before comparing them.", "why": "A bounded comparison can generate a useful local signal for a cheap next decision.", "caveat": "An informal comparison is not a randomized controlled experiment and does not establish statistical confidence, causality, or generalizability.", "gold_review": { "reviewed_at": "2026-09-15", "review_status": "gold-ready", "first_action": "Choose one reversible decision, write what you want to learn and the outcome you will inspect, then define versions A and B so one main difference is visible.", "observable_signal": "Whether the question and outcome were defined in advance, one main difference was interpretable, the observed result was recorded honestly, and the user can distinguish a tentative signal from stronger evidence.", "eligibility": [ "The decision is reversible or low-risk enough that trying two approaches does not create disproportionate harm.", "The user can state one clear question and one outcome that matters before observing the result.", "The comparison can keep major surrounding conditions reasonably similar or can explicitly record when they differ." ], "when_not_to_use": [ "Do not run informal experiments on people, production systems, health, safety, security, finances, legal obligations, privacy, or other consequential settings when the test itself can create material harm or requires approval.", "Do not call a small uncontrolled comparison a randomized A/B test or use it to claim statistical confidence, causality, or generalizability.", "Do not change the outcome measure after seeing which version looks better without clearly marking the analysis as exploratory." ], "evidence_boundary": { "status": "practical", "statement": "This is a practical compare-and-learn template. Brali does not claim that an informal two-version comparison establishes causality, statistical significance, a generalizable winner, or an evidence-backed optimal test design.", "source_decision_ids": [] } } } ], "answer_packet": { "schema_version": 1, "canonical_id": "brali:problem:decision-under-uncertainty", "problem": "How do I make a useful decision when I cannot know the outcome in advance?", "canonical_url": "https://brali-lifeos.github.io/problems/decision-under-uncertainty/", "best_fit": { "canonical_id": "brali:protocol:circles-of-control-planner", "slug": "circles-of-control-planner", "title": "Separate Outcomes From Actions You Can Control", "when": "The decision feels stuck because controllable actions and uncertain outcomes are mixed together.", "why": "Separating control, influence, monitoring, and uncertainty creates a cleaner action boundary before choosing a test or commitment.", "caveat": "The boundary can change over time and must not be used to dismiss structural constraints, risk, or legitimate escalation.", "first_action": "Write one concern, separate the uncertain outcome from your own observable actions, and choose the smallest useful behavior you can take without pretending it guarantees the result.", "evidence_state": "practical" }, "alternatives": [ { "canonical_id": "brali:protocol:start-with-a-hypothesis", "slug": "start-with-a-hypothesis", "title": "When Faced with a Problem, Start by Making an Educated Guess About the Cause", "when": "You have a plausible explanation for an unclear work problem and can check it with a small reversible observation or test.", "why": "A hypothesis plus a prediction makes disconfirming evidence visible and reduces random switching between theories and fixes.", "caveat": "Do not experiment casually in consequential systems and do not treat the first plausible explanation as a conclusion.", "first_action": "Write the observable problem separately from your explanation, then write one plausible hypothesis and one concrete pattern you would expect to see if it were useful.", "evidence_state": "practical" }, { "canonical_id": "brali:protocol:ab-test-learning-loop", "slug": "ab-test-learning-loop", "title": "Compare Two Approaches and Learn From the Result", "when": "Two approaches are both plausible, the choice is low-risk and reversible, and one observable outcome can be declared before comparing them.", "why": "A bounded comparison can generate a useful local signal for a cheap next decision.", "caveat": "An informal comparison is not a randomized controlled experiment and does not establish statistical confidence, causality, or generalizability.", "first_action": "Choose one reversible decision, write what you want to learn and the outcome you will inspect, then define versions A and B so one main difference is visible.", "evidence_state": "practical" } ], "stop_rule": "Do not run casual experiments when the test itself can create material health, safety, legal, financial, privacy, security, or production risk. Escalate the decision method when the stakes require stronger evidence." }, "evidence_decisions": [ { "id": "ai-review-correction-friction-boundary-2026", "decision": "propose-protocol", "reviewed_at": "2026-08-29", "source_title": "Bias in the Loop: How Humans Evaluate AI-Generated Suggestions", "source_url": "https://hdsr.mitpress.mit.edu/pub/nrcn4h7d/release/2", "supported_claim": "When humans review AI-generated suggestions, the review interface itself can bias behavior. In this experiment, adding repair work to the act of rejecting an AI suggestion reduced correction activity and increased undercorrection. Brali can therefore justify a bounded workflow rule: make it cheap to flag or reject an AI output, and separate validation from repair when the repair burden would otherwise make acceptance the path of least resistance.", "unsupported_or_overstated_claims": [ "Making correction easier will always increase overall accuracy.", "People who distrust AI are universally better reviewers.", "Performance bonuses cannot improve AI review in other settings.", "The same effect size applies to expert, medical, legal, financial, or safety-critical review.", "A human-in-the-loop label by itself guarantees reliable oversight.", "AI suggestions should be hidden from reviewers." ], "limitations": [ "The experiment used crowdworkers rather than domain experts.", "The task was limited to ten greenhouse-gas reporting tables and one pre-annotation workflow.", "Several difficult items required domain knowledge that many annotators lacked.", "The randomized correction-burden manipulation changed correction behavior, but the regression analysis did not show a clear overall accuracy loss because overcorrections also decreased.", "AI attitudes predicted behavior observationally rather than through randomized manipulation.", "The study did not include a no-AI baseline and could not analyze item-order effects." ] }, { "id": "ai-structured-intake-human-judgment-boundary-2026", "decision": "propose-protocol", "reviewed_at": "2026-08-29", "source_title": "Voice AI in Firms: A Natural Field Experiment on Automated Job Interviews", "source_url": "https://arxiv.org/html/2607.28222", "supported_claim": "A defensible way to divide some repetitive high-volume workflows is to automate structured information collection while keeping consequential evaluation with a human. In this field experiment, that division improved several downstream hiring outcomes without a measured decline in worker productivity. Transcript evidence is consistent with greater standardization and comparability as a mechanism, but does not prove that mechanism independently. Brali should treat this as a task-allocation pattern to test, not as evidence that AI should make final hiring or other high-stakes decisions.", "unsupported_or_overstated_claims": [ "AI interviewers are generally better than human interviewers.", "AI should make final hiring decisions.", "The result generalizes to specialized, relationship-heavy, tacit-knowledge, executive, clinical, legal, or other high-stakes work.", "Automating information collection removes discrimination or guarantees fairness.", "The same voice-AI system will produce the same results in other firms, languages, cultures, or labor markets.", "Human oversight automatically prevents automation bias.", "The controlled-variance mechanism is causally proven by the experiment." ], "limitations": [ "The source is a working paper rather than a peer-reviewed journal article.", "The experiment was conducted with one recruitment-process outsourcing firm and entry-level customer-service hiring in the Philippines.", "Five percent of AI interviews ended because applicants were unwilling to continue with AI and seven percent experienced technical failure.", "The transcript-based mechanism analysis is associative even though interviewer assignment was randomized.", "The candidate-experience survey had a low response rate and may not represent all applicants.", "Applicants who were allowed to choose showed negative sorting into AI, limiting interpretation of the choice condition.", "Employment selection has legal, fairness, accessibility, and accountability requirements that this study does not resolve." ] }, { "id": "consider-opposite-social-judgment-boundary-1984", "decision": "propose-protocol", "reviewed_at": "2026-09-11", "source_title": "Considering the opposite: A corrective strategy for social judgment", "source_url": "https://doi.org/10.1037/0022-3514.47.6.1231", "supported_claim": "When a judgment is vulnerable to one-sided evidence processing, deliberately generating an opposed possibility can reduce bias on some tasks more effectively than simply telling oneself to be fair or unbiased. Brali can use this as a concrete pre-decision check while preserving the possibility that the original conclusion remains correct.", "unsupported_or_overstated_claims": [ "Consider-the-opposite eliminates confirmation bias.", "The technique transfers automatically to every real-world decision.", "The opposite conclusion should be preferred once generated.", "More counterarguments are always better.", "The strategy should be used for every trivial or reversible choice." ], "limitations": [ "Classic laboratory/social-judgment evidence from undergraduate samples.", "Only two focal task domains were tested in the original article.", "Long-term persistence and broad transfer were not established.", "The authors note that considering the opposite can in some circumstances overweight disconfirming evidence and create a different form of partiality.", "Demand characteristics and task-specific effects remain possible." ] } ], "trust_note": "Recommendations come only from protocols that are both trusted in the Brali Protocol Feed and manually Gold-ready. The problem-to-protocol fit is editorial decision logic; it is not a claim that one sequence is universally best." }