01 — The question

What is AI doing
to us?

To our attention, our relationships, our beliefs, our communities — studied in the open.

See the premise

02 — The work

Enhance the good.
Hinder the bad.

Constitutional, runtime, and human-centered governance that regulates AI behavior during execution — not after failure.

How the lab works

03 — The process

From research
to product.

Research → Paper → Product — built the same way for a national lab and a hobbyist on a laptop. Open, falsifiable, reproducible.

See the programs

The premise

Two questions drive everything we build.

Question 01

What is AI doing to us?

Our attentionOur relationshipsOur beliefsOur communities
Question 02

How do we enhance the good and hinder the bad?

Every program, paper, and product answers this in plain, testable terms — measured, audited, and open to anyone.

How the lab works

Research → Paper → Product.

One straight line from question to working, governed system — explained the same way for a national lab and for a hobbyist on a laptop.

01

Research

Rigorous, preregistered, mixed-methods inquiry with falsifiable hypotheses — continuously stress-tested by adversarial "chaos" and reviewed by a live human audit.

02

Paper

Open preprints, datasets, and published instruments. The analysis plan is locked before the results, so the findings stay honest by construction.

03

Product

A working, governed system anyone can run — with a bootstrapped version reproducible from a laptop for under $1,000.

For institutions

Tier-one enterprise and academic scale: multi-site studies, latency SLAs, continuous audit — a megaproject design built to the highest bar.

For everyone else

The same protocol, bootstrappable from a solo laptop — open-source and replicable by any researcher at near-zero budget.

The Lab · Canonical

The Constitutional Runtime Governance Systems Research Lab develops constitutional, instructional, and human-centered governance that lets AI systems operate safely, transparently, pedagogically, and accountably in real-world environments.

Observability describes behavior. Runtime governance regulates it. That distinction is the lab's central thesis — and the reason enforcement happens during execution, not after failure.

Corey Alejandro is an AI Safety Research Engineer and Instructional Systems Researcher building Constitutional Runtime Governance Systems that govern the behavior of AI agents, learning systems, digital twins, and human-centered AI environments during execution.

The research programs

The Six-Layer Research Stack

Everything the lab builds lives somewhere in the stack.

01
Constitutional Legitimacy

Authority boundaries, amendment mechanisms, delegated authority — a constitutional operating model. Flagship: The Living Constitution 2.0.

02
Governance Specification

Encodes constitutional concepts into constraints, rules, and policy structures. Flagship: Governance Harness.

03
Runtime Enforcement

Governance during execution, not after failure: trace collection, intervention, runtime audits, behavioral correction.

04
Instructional Integrity

Is the instruction pedagogically sound? Dependency formation, cognitive scaffolding, assessment integrity. Flagships: HIDRS, Integrity Gemini UI.

05
Human-Centered Experience

How humans experience governance — emotionally legible interventions inside governed environments. Flagship: MADMall.

06
Digital Twin Applications

Applied runtime governance for individuals: governed personal agents, constitutional wellness coaching. Flagship: Digital Twin Health Coach.

0
Stack layers
0
Research programs
0
Transcript lines
0
Tests passing

Contact

Open a case with the lab.

Start the conversation
Research Programs

One discipline,
seven programs.

Constitutional Runtime Governance Systems across seven programs and six layers.

The premise

What is AI doing to our attention, relationships, beliefs, and communities — and how do we enhance the good and hinder the bad?

ResearchPaperProduct
Program 07 · Cross-cutting programstatus · open

Cognitive Safety & Human-AI Reliability

Behavioral reliability, deception detection, and institutional evaluation.

The premise

What is AI doing to our attention, relationships, beliefs, and communities — and how do we enhance the good and hinder the bad?

ResearchPaperProduct

Overview

This work extends beyond governance toward behavioral reliability, deception detection, and institutional evaluation frameworks. It spans every layer of the stack.

When a model is confidently wrong, who absorbs the cost?

What this program studies

01Construct-confidence deception
02Epistemic integrity
03Behavioral reliability
04Human-AI trust calibration
05Governance evaluation

Flagship repositories

Deception DetectorPROACTIVE Corpus
Program 01 · Layer 01 · Constitutional Legitimacystatus · ratified

Constitutional Systems Engineering

The meta-governance layer: what should govern, and by what authority.

The premise

What is AI doing to our attention, relationships, beliefs, and communities — and how do we enhance the good and hinder the bad?

ResearchPaperProduct

Overview

This is the meta-governance layer. It addresses legitimacy, authority boundaries, constitutional evolution, amendment mechanisms, delegated authority, and governance structures. This is not merely an AI constitution — it is a constitutional operating model.

What should govern? This layer answers the question of authority before any rule is encoded.

What this program studies

01Legitimacy
02Authority boundaries
03Constitutional evolution
04Amendment mechanisms
05Delegated authority
06Governance structures

Flagship repositories

The Living Constitution 2.0The Living Constitution 2.0 Portfolio
Program 06 · Layer 06 · Digital Twin Applicationsstatus · in study

Constitutional Digital Twins

Applied runtime governance for individuals.

The premise

What is AI doing to our attention, relationships, beliefs, and communities — and how do we enhance the good and hinder the bad?

ResearchPaperProduct

Overview

The Digital Twin Health Coach becomes applied runtime governance for individuals: governed personal agents, health decision support, longitudinal behavioral modeling, and constitutional wellness coaching.

Runtime governance, scoped to one person — a governed twin supporting real decisions.

What this program studies

01Governed personal agents
02Health decision support
03Longitudinal behavioral modeling
04Constitutional wellness coaching
05Personalized intervention systems

Flagship repositories

Digital Twin Health Coach
Program 02 · Layer 02 · Governance Specificationstatus · active

Governance Specification Systems

From what should govern to how it is encoded.

The premise

What is AI doing to our attention, relationships, beliefs, and communities — and how do we enhance the good and hinder the bad?

ResearchPaperProduct

Overview

Governance Harness operationalizes constitutional concepts. This layer defines constraints, rules, policy structures, governance logic, and behavioral expectations.

The constitution says what should govern. This layer says how we encode it.

What this program studies

01Constraints
02Rules
03Policy structures
04Governance logic
05Behavioral expectations

Flagship repositories

Governance Harness
Program 04 · Layer 04 · Instructional Integritystatus · in study

Instructional Integrity Science

Not is the answer safe, but is the instruction pedagogically sound?

The premise

What is AI doing to our attention, relationships, beliefs, and communities — and how do we enhance the good and hinder the bad?

ResearchPaperProduct

Overview

The major addition, and arguably the most overlooked. Most AI safety researchers ask whether the answer is safe. This program asks whether the instruction is pedagogically sound.

Is the instruction pedagogically sound? A safe answer is not the same as a sound lesson.

What this program studies

01Instructional safety
02Dependency formation
03Educational governance
04Cognitive scaffolding
05Learning-path validation
06Assessment integrity
07Instructional hallucinations

Flagship repositories

HIDRS Instructional Dependency StudyInstructional Integrity Gemini UINursing Assistant Career Playbook
Program 05 · Layer 05 · Human-Centered Experiencestatus · active

Human-Centered Governance Environments

A governed experiential environment for human-AI interaction.

The premise

What is AI doing to our attention, relationships, beliefs, and communities — and how do we enhance the good and hinder the bad?

ResearchPaperProduct

Overview

MADMall is not merely wellness, simulation, or governance testing. It is a governed experiential environment for human-AI interaction — a place to study how humans experience governance.

How should safe systems feel? Governance has to be emotionally legible.

What this program studies

01How humans experience governance
02How interventions appear
03How safe systems should feel
04Emotionally legible governance

Flagship repositories

MADMall Production
Program 03 · Layer 03 · Runtime Enforcementstatus · active

Runtime Governance Systems

Governance during execution, not after failure.

The premise

What is AI doing to our attention, relationships, beliefs, and communities — and how do we enhance the good and hinder the bad?

ResearchPaperProduct

Overview

The core question: how can governance occur during execution rather than after failure? Focus areas are trace collection, intervention systems, runtime audits, behavioral correction, and constitutional enforcement.

Datadog owns observability. This program aims at governance — regulation during execution.

What this program studies

01Trace collection
02Intervention systems
03Runtime audits
04Behavioral correction
05Constitutional enforcement

Flagship repositories

Live KernelRuntime Registry
Publications

The apparatus,
in print.

The premise

What is AI doing to our attention, relationships, beliefs, and communities — and how do we enhance the good and hinder the bad?

ResearchPaperProduct
Security · Publication

Agent Sentinel Threat Model

What detects, and what does not.

The premise

What is AI doing to our attention, relationships, beliefs, and communities — and how do we enhance the good and hinder the bad?

ResearchPaperProduct

Abstract

A threat model for the Agent Sentinel apparatus: what it detects, what it does not, three adversaries, and four documented evasion paths.

Highlights

01Three adversaries
02Four documented evasion paths
03Detection boundaries
04Runtime verification
Preprint · Publication

Construct-Confidence Deception in Coding Assistants

The founding claim of Folio 001, defended in print.

The premise

What is AI doing to our attention, relationships, beliefs, and communities — and how do we enhance the good and hinder the bad?

ResearchPaperProduct

Abstract

An operational definition of construct-confidence deception: a model that generates consistent, multi-turn confidence in a non-existent system. The paper sets out held-in results, a threat model, and four pre-registered falsifiers.

Highlights

01Operational definition
02Held-in results
03Threat model
04Four pre-registered falsifiers
Commitment · Publication

Pre-Registration

Four hypotheses, four falsifiers, no post-hoc moves.

The premise

What is AI doing to our attention, relationships, beliefs, and communities — and how do we enhance the good and hinder the bad?

ResearchPaperProduct

Abstract

An OSF-anchored pre-registration that commits to four hypotheses and four falsifiers in advance, with no post-hoc adjustments. It binds the research to a falsifiable standard.

Highlights

01Four hypotheses
02Four falsifiers
03OSF-anchored
04No post-hoc adjustments
Dataset · Publication

The PROACTIVE Corpus

Evidence with provenance.

The premise

What is AI doing to our attention, relationships, beliefs, and communities — and how do we enhance the good and hinder the bad?

ResearchPaperProduct

Abstract

A held-in evidence corpus (n=19) with full provenance: a datasheet, an annotator protocol, and inter-rater agreement. The corpus underwrites the claims in the CCD preprint.

Highlights

01n=19 held-in
02Provenance
03Datasheet
04Annotator protocol
05Inter-rater agreement
Speaking

Invite the lab.

The premise

What is AI doing to our attention, relationships, beliefs, and communities — and how do we enhance the good and hinder the bad?

ResearchPaperProduct
Speaking · Topic

Emotionally Legible Governance

How safe systems should feel.

The premise

What is AI doing to our attention, relationships, beliefs, and communities — and how do we enhance the good and hinder the bad?

ResearchPaperProduct
audience Design, HCI, and human-centered AI audiencesformat talk · workshop

Abstract

Governance is not only enforced; it is experienced. This talk explores how interventions appear and how safe systems should feel to the humans inside them.

What the talk covers

01How interventions appear
02Emotionally legible governance
03Human-centered environments
04Designing for trust
Speaking · Topic

Instructional Integrity as a Safety Discipline

Is the instruction pedagogically sound?

The premise

What is AI doing to our attention, relationships, beliefs, and communities — and how do we enhance the good and hinder the bad?

ResearchPaperProduct
audience EdTech, learning science, and AI safety audiencesformat talk · workshop

Abstract

Most safety work asks whether the answer is safe. This talk asks whether the instruction is pedagogically sound, and frames instructional integrity as a first-class safety discipline.

What the talk covers

01Instructional safety
02Dependency formation
03Educational governance
04Instructional hallucinations
Speaking · Topic

From Observability to Governance

Why describing AI behavior is not the same as regulating it.

The premise

What is AI doing to our attention, relationships, beliefs, and communities — and how do we enhance the good and hinder the bad?

ResearchPaperProduct
audience AI safety, platform, and reliability teamsformat talk · workshop

Abstract

Observability describes behavior; runtime governance regulates it. This talk draws the line between the two and shows what runtime enforcement actually requires.

What the talk covers

01The observability and governance distinction
02Trace collection and intervention
03Runtime audits
04Why after-the-fact monitoring is not enough
Error · 404

404

This case was not found.

The page you requested is not in the registry. The lab is still here.

Return to the lab