Node & Norm studies the evidence behind AI-assisted decisions: how they are evaluated, authorized, challenged, and corrected. The research examines whether safeguards worked and whether the evidence exists to show it.
Explore the research · Where the work stands · Brand assets
Policies, risk reviews, and approvals establish expectations for AI use. A consequential decision still needs evidence that those expectations held. What supported the recommendation? Who could challenge it? Did the correction reach execution?
Node & Norm develops methods, evidence records, and reference implementations for examining those questions. Sources, versions, and unresolved findings stay attached to the work.
Can the person assigned responsibility actually affect what happens?
| Research | What to inspect |
|---|---|
| Trust, Autonomy & Evidence (TAE) | Conditions for practical human control, from evidence access to an intervention taking effect. Three selected historical cases; one assessor. |
| Human Influence Telemetry (HIT) | What documentary records establish about the exercise of human authority. Broader validation remains open. |
Can a team reconstruct why it proceeded, stopped, or waited, and what obligations remained?
| Research | What to inspect |
|---|---|
| Governed Decision Intelligence (GDI) | General decision-record specification and reference implementation. |
| AI Assistance Governance | A method/profile for bounded AI participation, assigned human authority, and dependency assessment. |
| Regulated Gate Decision Support (RGDS) | Regulated phase-gate reference implementation, illustrative records, and validation tools. |
| RGDS independent study | Historical, exploratory research with its own record. Read the study. |
GDI supplies the general decision architecture, AI Assistance Governance defines its participation profile, and RGDS supplies a regulated reference implementation. Availability of code does not establish field effectiveness.
Do evaluation rules, implementation, and evidence support the reported result?
Catholic Doctrinal Fidelity Index (CDFI) develops evaluation-governance methodology and a reference implementation for domain-specific doctrinal assessment. Scoring-contract reconciliation remains open. SAICRED is an external collaborative project with its own leadership and publication authority.
These are different dimensions: publication status, ongoing research, and source availability. Status reflects the author's website update of 15 September 2026; it does not advance the scope of earlier findings.
| Work | Dimension | Current record |
|---|---|---|
| TAE manuscript | Publication status | On hold pending arXiv approval |
| HIT | Research status | Research in progress |
| CDFI | Research status | Evolving |
| GDI | Availability | Open source |
| RGDS | Availability | Open source |
| Poenitentia Institutionum | Publication status | Editorial review |
Emerging research and shared infrastructure
AI Incident Reconstruction & Assurance examines what happened, which controls held, and what supports a correction. Website case records remain provisional documentary reviews.
Institutional Responsibility & Repair asks what an institution owes those affected by AI-mediated harm. Poenitentia Institutionum develops the normative argument; its repository is private and empirical outcome validation remains open.
Control Evidence Corpus (CEC) is private development infrastructure for testing whether independent reviewers can reconstruct control operation from surviving evidence. Its empirical integration with downstream projects remains prospective.
These areas preserve the three permanent research programs above. The website Registry owns presentation identities and admitted summaries. Each source repository retains its research and release authority.
Research workflow, provenance, and reuse
Applied AI Research Translator turns research inputs into traceable claims, bounded tasks, and decision records subject to human review.
Each project maintains its own scope, versions, validation status, and releases. Reuse must preserve source dependence, missingness, evidence cutoffs, and exact-version provenance. Shared organizational ownership supplies no additional validity to a method or result.
Start with the project's research-status statement, then inspect its versioned release, sources, and validation instructions. Use its contribution and correction process. Cite the authors and exact artifact version.
Founded by Mark Julius Banasihan. Independent replication and field studies remain necessary to establish broader reliability and practical benefit.
Every decision is a Node. Every standard is a Norm.