Sign In
Sign in to access your HOM assessments.
Assessments
Structured human-outcome assessment and causal pathway analysis.
New Assessment
Assessment Overview
Practical HOM · AI Pipeline
Development checkpoint: bounded AI operations with deterministic validation and saved checkpoints. Existing Modules A–F are unchanged.
P1 · Guided, Open Intake
Paste source material or a concise description. A focused question is optional. This first bounded operation summarizes the material neutrally, surfaces possible issues, and keeps the inquiry open rather than deciding the question for the user.
P1A · Validated Intake Checkpoint
Only structurally and semantically valid intake output is saved.
P2 · Focused Question & Claim/Evidence Setup
After intake, enter the question you want HOM to examine. This bounded operation defines scope, decomposes only materially relevant claims, and separates what the supplied source establishes from what requires external evidence. It does not research or answer the question.
P2A · Validated Focused Checkpoint
A valid P1 intake checkpoint is required. Only validated focused output is saved.
P2B · Focused Validation Record
P3 · External Evidence Retrieval
Uses browser search only after a validated P2 checkpoint exists. This stage retrieves potentially relevant external sources; it does not validate those sources, answer the question, or reach a HOM conclusion.
P3A · Retrieved Research Checkpoint
A saved retrieval checkpoint records candidate sources and traceable URLs. Saving it does not mean the sources or claims have passed HOM evidence validation.
P3B · Retrieval Validation Record
P4 · Source & Evidence Validation
Validates candidate sources one at a time in two bounded stages. Source verification is saved before evidence extraction, so a provider cooldown never repeats a completed verification. A failure never regenerates or discards work that already passed.
P4A · Source-by-Source Validation Checkpoints
Each passed source is saved independently. Source quality and applicability remain separate; verification does not establish a HOM conclusion.
P4B · Source Validation Record
A — Decision & Alternatives
Define the decision, feasible alternatives, and the explicit reference counterfactual used for comparison.
A1 · Decision Definition
A2 · Alternatives
Alternatives are candidate courses of action. The reference counterfactual is defined separately below.
A3 · Reference Counterfactual
Describe the most plausible comparison world if the decision is not implemented. “No action” does not necessarily mean “no change.”
B — People & World-State Change
Define who may be affected, what changes in the world, and which populations are exposed to those changes.
B1 · Affected Populations
Identify materially relevant populations and make the assessment boundary explicit.
B2 · World-State Changes
Record intermediate changes in conditions. A world-state change is not automatically a human outcome.
B3 · Population–Change Relationships
Connect populations to the world-state changes they may experience. This establishes exposure; it does not yet assert a HOM outcome.
C — HOM Screening
Screen each population–change relationship separately for plausible favorable and adverse material pathways.
D — Pathways & Evidence
Build explicit causal pathways from screened population–change relationships to HOM outcomes. D1 creates pathway records. D2 adds ordered causal nodes. D3 records explicit causal claims between nodes. D4 attaches evidence to specific causal links.
D0 · Module C Screening Disposition
Trace each eligible Module C screening result into Module D. Existing D1 pathways are recognized as Pathway Developed.
D1 · Causal Pathways
Create a pathway only where a plausible or uncertain material pathway was identified in Module C.
D2 · Pathway Nodes
Build the ordered causal states within a selected pathway. Nodes describe states or mechanisms; causal claims between them will be recorded as links in D3.
D3 · Causal Links
Record the causal claim represented by each arrow between pathway nodes. Evidence will attach to these links in D4.
D4 · Evidence
Attach evidence to a specific causal link. Adding evidence does not automatically change the link support state.
E — Human-Outcome & Population Comparison
Characterize valid Human Outcomes from Module D and compare how outcomes are distributed across affected populations. Module E does not replace or repeat the causal analysis in Module D.
E1 · Outcome Register
Valid Human Outcome endpoints are populated automatically from Module D.
E2 · Outcome Characterization
Characterize each valid Human Outcome for expected magnitude, population reach, duration, and confidence.
E3/E4 · Population Comparison, Distribution & Tradeoffs
A comparison is required only where two or more valid Human Outcomes occur within the same HOM dimension.
F — Assessment Synthesis
Synthesize the completed HOM analysis without collapsing distinct human outcomes into a numerical score.
F1 · Assessment Scope & Coverage
Current assessment coverage is generated automatically from Modules A–E.
F2 · Human-Outcome Synthesis
Synthesize the current valid Human Outcomes within each represented HOM dimension. A saved synthesis requires review if its outcome membership changes.
F3 · Material Uncertainties & Tradeoffs
Identify assessment-wide issues that materially affect interpretation across dimensions or populations.
F4 · Overall Assessment Conclusion
Record the analyst's overall HOM finding after considering the complete assessment. The finding is not calculated automatically.
HOM Assessment Report
Read-only report generated from the current A–F assessment state.