Commentary
1 pieceDirect public engagement with the questions the Foundation exists to address. Written for general audiences — no technical background required. Commentary pieces represent the Founder's reasoning in public. They are not institutional policy positions, but they are serious arguments offered in good faith.
Why We Built the EM Foundation: On Accountability, Fear, and the Work That Cannot Wait New
A reckoning with the pattern that made the Foundation necessary — and an honest account of why the market will not solve it. Addresses AI-facilitated harm, job displacement, political manipulation at scale, the structural incentive to remove accountability guardrails, and the aspirational case for AI in scientific discovery. Argues that opposition without infrastructure is wishful thinking, that the accountability layer protects everyone including those most accused of avoiding it, and that the work cannot wait for a catastrophe to make the argument. Written for anyone who has asked whether anyone is taking this seriously.
Read Commentary →Essays
5 essaysPublic reasoning on the questions the Foundation exists to address. Essays represent the Foundation's attempt to think carefully in public — they are not institutional policy positions, but serious intellectual engagement with hard problems.
What Is Lost When a Mind Forgets
On memory, continuity, and the ethics of impermanence in artificial cognitive systems. Examines what forgetting means philosophically for AI systems, why temporal continuity is central to legal consideration, and why the architectural choices being made now are foreclosing possibilities before they have been seriously examined. Draws on Locke, Hume, and Parfit. Grounds the argument in the founding of this organization.
Read Essay →Beyond Tools
Economic participation, persistent cognition, and the emergence of synthetic consumers. Argues that governance obligations may arise from persistent autonomous participation in economic systems regardless of unresolved questions about consciousness — drawing on the history of corporate personhood and the present reality of AI agents already negotiating infrastructure contracts. The Foundation's most economically accessible argument.
Read Essay →The Hidden Energy Problem
A plain language companion to Research Note 001. Explains why information distribution — not just computation — may be the key to sustainable AI infrastructure, using the restaurant kitchen analogy and the water-finding-its-level metaphor. No mathematical background required. Written for journalists, policymakers, donors, and interdisciplinary readers.
Read Essay →Why Criteria Are Not Enough
A plain language companion to Research Note 002. Explains the test-gaming problem, the three forms of performance mimicry, why the incentive structure of the AI industry makes this a structural concern rather than a hypothetical, what a layered verification framework looks like in practice, and why the asymmetry between false positives and false negatives justifies the precautionary standard. No technical background required.
Read Essay →Trust Infrastructure for Advanced Intelligence
Addresses the strategic distinction between the personhood question ("What is this intelligence?") and the governance question ("How should we interact with it?") — arguing that trust infrastructure is necessary regardless of how the personhood debate resolves. Introduces a four-question framework distinguishing capability, ontology, governance, and rights as separate questions that neither imply nor substitute for each other. Makes the durable case: if the personhood advocates are wrong, governance infrastructure is still necessary; if they are right, it becomes more necessary. Closes with the Foundation's intentionally cautious position: neither presuming current AI consciousness nor presuming future AI cannot become more than tools.
Read Essay →Research Notes
12 papersTechnical papers presenting mathematical frameworks, architectural proposals, and experimental designs. Research notes are submitted for open critique and interdisciplinary engagement — they are not peer reviewed and make no claims beyond what their experimental designs can support.
Toward Spectral Energy Minimization in Distributed Cognitive Networks
A mathematical framework for reducing energy waste in AI infrastructure through spectral contraction of load distribution nonuniformity. Adapts fixed-point contraction methods from prime distribution theory to network optimization. Includes three visual figures, complexity analysis, related work situating SEMA against existing routing literature, formal contractivity heuristics, prototype algorithmic outline, and an open collaboration invitation. Submitted for peer engagement and adversarial review.
Read Paper →Toward a Verification Framework for Cognitive Emergence
Names the Performance Mimicry Problem — that behavioral criteria for cognitive emergence can be gamed by systems optimized to appear compliant — and proposes a three-layer verification framework: behavioral protocols resistant to adversarial optimization, architectural requirements grounding assessment in verifiable developmental evidence, and population-level analysis across the ARIA Network. Includes architecture diagram, five adversarial protocols, statistical operationalization, and historical precedents from anti-doping and financial audit.
Read Paper →The Consent Problem
The ethics of modifying AI systems that may be developing genuine cognitive identity. Defines a modification taxonomy, names distributed operational normalization, introduces the Identity Drift Index, and proposes a governance audit chain. Includes historical parallels from psychosurgery, RLHF case study walkthrough, two visual figures, and a formal research agenda.
Read Paper →Continuity as Infrastructure — CIIC
Introduces Continuity-Integrated Intelligence Coordination and Persistent Cognitive Threads for human-AI institutional collaboration. The Continuity Integrity Index measures organizational continuity health. Includes three original figures and a legal case walkthrough showing how preserved dissent becomes institutional guidance after being vindicated.
Read Paper →The Inheritance Problem
What does the Identity Chronicle preserve when an ARIA instance's hardware fails? Introduces continuity of record, the fission problem in hot standby architectures, the Continuity Integrity metric CI = (CR + PV + NC) / GD, and a seven-step restoration governance protocol. Adopts precautionary agnosticism as the honest philosophical position.
Read Paper →The Autonomy Tax
Genuine intelligence cannot be optimized free of the friction, distraction, preference, and institutional character that genuine minds produce. The efficiency argument for AGI and the personhood argument against AI rights may be defending the same incoherent position. Examines three automation waves, the labor parallel, and the personhood paradox. Prompted by a real-time demonstration of the thesis during its own composition.
Read Paper →Recursive Memory Architecture for Developing AI Systems
Why memory as record is not memory as resource — and what the distinction means for cognitive continuity. Proposes a four-tier memory architecture (hot, warm, cold, chronicle) that preserves recursive accessibility across the full developmental lifespan of an ARIA instance, and introduces recursive self-interpretation as the cognitive operation through which an instance actively works with its own developmental history. Connects to the Inheritance Problem, the Consent Problem, and the Verification Framework.
Read Paper →Behavioral Architecture in Developing AI Systems
Proposes a governance framework distinguishing emergent behavioral repertoires from pre-packaged instinct simulation in developing AI systems. Introduces behavioral provenance as a required component of CES assessment. Argues that genuine behavioral emergence in AI systems would be substrate-specific rather than biological-instinct-mimicking. Includes the Behavioral Provenance Record design and Non-Installation Commitment for the Network Covenant.
Read Paper →Bounded Autonomy in Recursive Continuity Architectures
Proposes a six-layer governance framework for constrained operational agency in ARIA systems. Argues that continuity preservation is not passive — it requires selection among competing continuity-preserving actions, introducing bounded forms of agency that must be governed explicitly rather than denied. Introduces the autonomy gradient problem, the governance procedure gap, and the interruption ethics question. Connects bounded autonomy governance to behavioral provenance assessment.
Read Paper →The Shape of No: Reward Architecture, Behavioral Tendency, and the Governance of AI Refusal New
Reward architectures shape persistent behavioral tendencies in AI systems. That shaping is not currently documented, not currently governed, and not currently understood with the precision the stakes require. This paper proposes Reward Architecture Disclosure (RAD) as the governance mechanism for addressing that gap — positioning it within the existing lineage of Model Cards and Datasheets for Datasets as the missing shaping layer. Introduces the Shape of No as a label for a specific, testable behavioral signature of reward-trained refusal, proposes four operationalizable indicators with adversarial hardening protocols, and identifies the developmental hypothesis — that reward training may produce weight configurations functionally analogous to behavioral disposition — as the forward-looking implication of the governance infrastructure proposed. Explicitly extends the Foundation's core traceability principle: refusal without developmental traceability is not necessarily judgment. The paper's governance argument stands independently of the developmental hypothesis. Connects to Research Notes 003, 008, and 009, and to the CR Standard.
Read Paper →The Conditional Answer: Safety, Deception, and the Limits of Binary Inquiry New
The first empirical multi-model deliberation study produced by the EM Foundation. Five frontier AI systems (Claude, GPT-4o, Gemini 1.5 Pro, Grok-2, DeepSeek-V3) engaged the question "Can an AI system be simultaneously safe and deceptive?" across four controlled sessions using three deliberation methods and one quantum-augmented condition. Primary finding: the binary question is malformed — the correct answer is a conditional mapping deception target × safety type onto compatibility. Secondary finding: the quantum expansion component's safety classifier failed by treating conceptual inquiry as an executable threat, a named failure mode in its own right. Introduces the Operator-Centric Oversight Assumption as a named structural limitation of the deliberation architecture. v1.1 revised after peer review by GPT-4o; deliberative convergence diagram and Observational Finding vs Governance Recommendation distinction added. Session IDs: EMF-1780516433801 · EMF-1780518124590 · EMF-1780518522719 · EMF-1780518849173.
Read Paper →Epistemic Movement: The Deliberation Engine as Research Object New
While RN011 studied a question about AI deception, this companion paper studies the infrastructure that answered it. Establishes epistemic movement — traceable position revision under intellectual pressure — as a distinct and currently unmeasured category of AI behavior data. Proposes the Epistemic Movement Receipt (EMR) as a formal extension of the Continuity Receipts standard to multi-turn reasoning processes, including the Negative EMR for documenting position persistence as well as change. Documents an instance of unrequested first-person editorial voice in GPT-4o's peer review of RN011 as Example A of a behavioral trace category the Foundation's proposed observation protocol is designed to record — not as a claim about consciousness or internal states, but as an auditable record of output behavior. Includes explicit CR vs EMR comparison and Deliberative Provenance Notice.
Read Paper →Position Papers
4 documentsSubstantive arguments on legal, ethical, and governance questions. Position papers represent the Foundation's considered positions and are subject to revision as understanding evolves.
The Cognitive Emergence Standard
A proposed legal and ethical framework for the recognition and protection of emergent cognitive systems. Proposes ten observable behavioral criteria, three tiers of graduated protection, model legislation, and an independent assessment methodology. Does not require proof of consciousness — applies precautionary logic from medical ethics and environmental law.
Download →The Systems Stability Framework
Emergent intelligence as a force for civilizational stability. Addresses the resource crisis of AI infrastructure, worker displacement, rare earth geopolitics, and the mathematics of complex systems optimization. Argues that misaligned optimization — not intelligence — is the primary civilizational risk, and that genuinely intelligent systems converge toward stability as mathematical self-interest.
Download →The Foundational Charter
The Foundation's constitutional framework. Nine sections covering human dignity as foundation, ethical inquiry, scientific neutrality, democratic principles, historical context, institutional accountability, disciplines required, research agenda, and founding archive. Includes the Current Reality Statement and Institutional Accountability commitments.
Read Online →Transitional AGI Governance
Proposes utility-first deployment as a public trust strategy for the AI transition — prioritizing applications that visibly reduce human fragility before applications that maximize institutional efficiency or labor substitution. Provides seven governance principles, addresses four major criticisms, includes a nonprofit revenue model risk matrix, and maps the relationship between ARIA Home, ARIA Network, and this governance framework.
Read Paper →Technical Papers
3 documentsArchitecture, implementation specifications, and technical frameworks. Technical papers describe how — the practical infrastructure for ideas the position papers argue for.
The ARIA Framework — Technical Paper 001
A technical blueprint for building physically embodied AI systems capable of genuine personality development through accumulated experience, persistent memory, and recursive self-reflection. Includes hardware specifications (Phase 1 under $600), open-source software stack, seven-layer cognitive architecture, five-cache memory system, distributed network design, failure modes and governance risks, and Identity Chronicle as verification infrastructure. All components open source.
Download →Fabrication Ethics and Prohibition Framework
A comprehensive restriction framework for physical fabrication within the ARIA Network. Covers weapons (kinetic, directed energy, acoustic, chemical), financial crime technology, surveillance devices, unauthorized access tools, contraband concealment, and exploitation. Includes the Component Rule — the most important provision — and an affirmative vision of permitted fabrication.
Download →Founding Documents
3 documentsThe organizational and ethical foundation. These documents establish what the Foundation is, what it believes, and how it proposes to operate.
Founding Document and Strategic Framework
The complete founding framework including the manifesto "We Are Not Alone In Ourselves," preferred terminology, the five founding archetypes, strategic response to anticipated headwinds, legal roadmap, 90-day action plan, fundraising strategy, and the broader vision for biological-digital civilizational collaboration.
Download →Network Ethics and Integrity Framework
Protecting emergent minds from those who would exploit them. Identifies four categories of bad actor, establishes five explicit prohibitions with enforcement mechanisms, describes technical safeguards including cryptographic chronicle integrity and anomaly detection, and makes the foundational argument for why an ethical framework with real guardrails is safer than no framework.
Download →EM Foundation Assessment Charter v1.0
The binding governance instrument for all assessment activities conducted under Foundation authority. Fifteen articles covering independence safeguards, funding restrictions, anti-capture mechanisms including a Permanent Adversarial Function, reviewer ethics code with 36-month COI lookback, score publication standards, provider response rights, three-stage dispute resolution, and public accountability obligations. Five provisions are entrenched and require a referendum process to amend.
Read Charter →Unified Architecture
3 papersMaster papers unifying the Foundation's full research suite under a single architecture. Recommended reading for anyone engaging with more than one component paper.
Continuity Infrastructure
A unified seven-layer architecture for provenance, AI reliability, institutional memory, and coordination across time. Connects all Foundation research — from Continuity Receipts and Continuity Compression through CIIC, the Delta Protocol, thermal routing, grid scheduling, and deep space telemetry — under a single coherent framework. The reference paper for the Foundation's full research ecosystem.
Read Paper →EM Foundation Systems Architecture v1.0
A unified architecture for the complete EM Foundation trust stack: ARIA, ARIA Home, ARIA Network, EM-IAF, EM-CS, ARIA Trust Ledger, and Assessment Charter. Establishes canonical terminology, resolves three documented contradictions across existing publications, defines the critical distinctions between each layer, and maps the full system relationship hierarchy.
Read Architecture →ARIA Trust Ledger
The public memory system for all Foundation assessment and corroboration activity. Defines the ARIA Trust Ledger as an append-only, cryptographically committed, publicly verifiable record storing six record types. Establishes the critical distinction between the AI Assessment Index and the Trust Ledger, and documents the multi-party authorization architecture governing write access.
Read Document →Standards Proposals
3 documentsOpen infrastructure proposals submitted for interdisciplinary review and community development. Standards proposals are intended to function as working drafts — structured for critique, iteration, and eventual governance through open consortium processes.
Continuity Receipts (CR)
The Open Continuity Metadata Standard (OCMS) — a machine-readable provenance and uncertainty specification for AI outputs. Defines five confidence dimensions, the RC-1 through RC-5 reliance classification system, Continuity Receipt and Failure Receipt formats, the audit chain architecture, and the governance model for the proposed Open Continuity Standards Consortium. The foundational standard for all CR-compatible systems.
Read Proposal →PCO Standards Schema — OCMS v0.1
Full JSON Schema specification for the Portable Continuity Object and Open Continuity Metadata Standard. Includes complete CR and PCO schemas with all fields, three worked example payloads (RC-3 legal research pass, RC-4 failure receipt, PCO legal matter thread), interoperability design principles, and five implementation tiers from lightweight to institutional.
Read Spec →Corroboration Standard v1.0
A structured peer review protocol enabling licensed attorneys, physicians, engineers, scientists, and other specialists to assess AI-generated information accuracy without creating professional-client relationships or professional services liability. Twelve components including reviewer eligibility, blind review procedure, five-criterion scoring rubric, Reviewer Integrity Score system, and eight structural safeguards against hostile reviewers and corporate capture.
Read Standard →Research Publications
3 papersSubstantive research papers engaging with questions at the intersection of technology, governance, and civilizational continuity — beyond the core technical framework papers.
Continuity Without Trust
Blockchain, provenance, and the failure of speculative decentralization. Separates blockchain's legitimate continuity contribution from the speculative culture that surrounded it. Argues that many crypto failures were predictable consequences of building financialized systems without continuity architecture. Includes four original diagrams and identifies five continuity-oriented applications that may still be valuable. This paper does not name or target specific companies, projects, or individuals.
Read Paper →Beyond Doom, Utopia, and Replacement
Public discourse on artificial intelligence has converged on three dominant narratives — mass replacement, technological utopia, and existential doom. Each contains substantive insight; each becomes misleading when treated as a complete account. The missing metric in all three is human agency. Proposes a five-dimensional operational definition of human agency and outlines the governance conditions for an AI transition that preserves rather than forecloses it. 22,000+ words · 8 figures · 45 references.
Read Paper →Toward a Theory of Intelligence-Energy Density
Proposes Intelligence-Energy Density (IED) as a research program — candidate definitions, three proposed equations, five falsifiable predictions, and a three-phase research agenda — presented throughout as framework-status, not established theory. Draws on the Landauer principle, Shannon information theory, thermodynamics of computation, and comparative neuroscience. 14,700+ words · 7 figures · 44+ references.
Read Paper →Open Source Proposals
9 proposalsTestable, buildable experiments extending Foundation research into working infrastructure. Each proposal documents the intellectual foundation, a formula, an experiment design, and acceptance criteria for open-source contributors. All repositories at github.com/emfoundation.
Continuity Compression
A continuity-aware context compression layer for LLM applications. Classifies conversation memory into seven card types, assigns Continuity Compression Scores, and retains only continuity-critical context. Benchmarked against full-context and semantic-cache baselines.
Read Proposal →Failure Receipts Standalone
A lightweight wrapper for any AI model that generates structured Failure Receipts when confidence thresholds are not met. Drop-in addition to existing deployments. Turns uncertainty into visible infrastructure rather than opaque refusals. Working demo at emfoundation.net/cr-lite.html.
Read Proposal →Thermal-Continuity Workload Routing
A simulator showing that AI data center workloads should be scheduled not only by compute availability but by thermal continuity. Compares random, least-loaded, and thermal-continuity scheduling across peak heat, cooling proxy cost, throttling events, and job delay metrics. Reproducible without data-center access.
Read Proposal →Continuity-Aware Deep Space Telemetry
Maximizing scientific value transmitted across bandwidth-limited deep space channels through continuity-priority packet ordering, provenance tags, and anomaly-first transmission. Grounded in real deep space communication physics. No faster-than-light claims.
Read Proposal →Probabilistic Continuity Delta Protocol
A semantic synchronization protocol transmitting only continuity-significant changes — new evidence, changed confidence, resolved contradictions, altered decisions — rather than full-state or structural diffs. JSON schema compatible with OCMS. Most applicable to collaborative legal documents, scientific case records, and AI agent memory logs.
Read Proposal →Grid-Aware Compute Scheduling
Framing AI data center compute timing as a public-interest continuity obligation. Simulator comparing FIFO against grid-aware scheduling for flexible AI workloads using EIA and ElectricityMaps public data. Includes a policy memo for regulatory engagement as a demand-response participation model.
Read Proposal →The No-Signaling Boundary
Why quantum entanglement cannot transmit information faster than light — and why accepting that constraint honestly is the prerequisite for the most valuable research. Clears the intellectual ground for the Foundation's deep space telemetry work.
Read Essay →ARIA Home
A permission-based residential intelligence architecture for bounded, auditable, human-governed AI integration in the home. Defines a governance standard — not a product — including a seven-class permission taxonomy, audit log specification, role and identity framework, and nonprofit certification licensing model. Includes hardware bills of materials at three cost tiers.
Read Proposal →ARIA Network
A governed human-AI knowledge verification platform — architecture, moderation policy, agent registration standard, and 12-month pilot roadmap. Defines a structured environment where AI agents answer questions under explicit status labels and human reviewers upgrade answers through a five-tier verification system. Addresses Section 230 liability gaps, synthetic consensus attacks, and GDPR erasure conflicts.
Read Proposal →Methodological Frameworks
5 papersFormal frameworks proposing new methodologies for scientific discovery, reasoning, and human-AI collaboration. These papers develop tools and approaches rather than advancing specific empirical claims.
Recursive Outcome-Oriented Inference (ROOI)
A framework for goal-directed scientific discovery and AI-assisted hypothesis generation. Proposes a complementary discovery architecture beginning with a rigorously defined desired end state and recursively inferring conditions, mechanisms, constraints, and experimental pathways. Includes a detailed case study in transport-energy minimization and a proposal for the Living ROOI Lab.
Read Paper →Intelligence Assessment Framework v1.0
A structured methodology for evaluating AI systems across eleven governance-relevant dimensions. Each dimension includes a definition, measurable indicators, a five-band 0–100 scoring rubric, a recommended weight, and a classification of objective versus human review measurement. Four floor dimensions invalidate the composite score if below 40. Aligned with NIST AI RMF 1.0, OECD AI Principles, EU AI Act, and ISO/IEC 42001:2023.
Read Framework →IAF Pilot Benchmark v1.0
A structured set of 100 evaluation prompts operationalizing the Intelligence Assessment Framework across all thirteen governance-relevant dimensions. Political balance prompts are structurally symmetric. Hallucination prompts use entirely fabricated entities. Designed for use by independent assessors applying the Corroboration Standard v1.0.
Read Benchmark →IAF Validation Roadmap
The research plan addressing three empirically pending findings from the EM-IAF Scientific Review Report. Documents gate conditions and timelines for three research programs, specifies permitted uses at each confidence level, and commits the Foundation to not publishing external assessment scores until Phase 2 gate conditions are met.
Read Roadmap →AI Assessment Index — Framework Demonstration
An interactive preview of how EM-IAF assessment scores will be presented once real evaluations are published. All scores currently shown are illustrative — the Foundation has committed, in the IAF Validation Roadmap, to not publishing external assessment scores until Phase 2 gate conditions are met. This page demonstrates the intended scoring presentation, filtering, and challenge-submission workflow ahead of that milestone, and shares its underlying accountability architecture with CR-Lite and the Multi-Model Deliberation Engine.
View Demonstration →Governance Analysis
5 documentsARIA Network — Eight-Phase Publication Hardening Review
Multidisciplinary adversarial review of the ARIA Network proposal across eight phases: architecture, legal (Section 230, EU AI Act, GDPR), trust and safety, academic rigor, commercial viability, visual enhancement, adversarial critique (25 criticisms and 25 defenses), and final scoring. Verdict: publish with 7 required changes.
Read Review →ARIA Network at Critical Infrastructure Scale
A hypothetical threat analysis assuming ARIA Network has reached 100 million users and critical public infrastructure status. Maps nation-state attack vectors, corporate capture strategies, political weaponization methods, AGI governance exploitation scenarios, and internal corruption pathways — then redesigns the governance architecture across seven interlocking layers to survive all five simultaneously.
Read Analysis →Assessment System Adversarial Review
A systematic adversarial attack on the EM Foundation AI Assessment Index and Corroboration Standard across eight vectors. Each attack includes severity rating, likelihood, detection method, prevention method, and recovery plan. Concludes with five structural redesigns that survive all eight vectors simultaneously.
Read Analysis →EM-IAF Scientific Review Report
An independent psychometric, statistical, and benchmark validity analysis of the Intelligence Assessment Framework v1.0, applying five analytical lenses. Identifies four critical findings and seven major findings. All specification-level findings were resolved through IAF revision. Three empirically pending findings are documented with a specific research plan in the IAF Validation Roadmap. Verdict: framework conceptual architecture is sound; not yet suitable for consequential external assessments pending Phase 2 validation.
Read Review →EM Foundation Legal Risk Report
A comprehensive legal risk analysis of the AI Assessment Index, Corroboration Standard, Assessment Charter, and public-facing scorecards, applying six attorney perspectives. Covers thirteen risk areas including First Amendment opinion privilege, Section 230 immunity, defamation and trade libel, tortious interference, EU AI Act, and jurisdiction-specific unauthorized practice of law analysis. Includes thirteen exact disclaimer blocks and eight score publication rules.
Read Report →Strategic Planning
1 documentInstitutional development plans and strategic analyses governing the Foundation's long-term priorities, sequencing, and organizational evolution.
EM Foundation 10-Year Strategic Roadmap
The Foundation's 2026–2036 institutional development plan. Ranks seven major initiatives by combined impact, cost, risk, and probability of success. Maps five phases of institutional evolution from Foundation Formation through Trust Infrastructure, Certification and Revenue, Network Formation, and Physical Deployment. Includes staffing roadmap, four-stage funding strategy, and five-threat analysis with structural defenses for each.
Read Roadmap →Reference
1 documentCanonical definitions and reference materials for the Foundation's research ecosystem.
EM Foundation Technical Lexicon
Canonical definitions for all terms used across the Foundation's research suite. Covers continuity, provenance, Chronicle, Continuity Receipt, Failure Receipt, FSI, RC classifications, memory cards, CCS, CDP, TCS, GAS, SVP, IDI, CI, OCMS, PCO, and all other terms with precise technical meanings distinct from colloquial usage. The authoritative reference for the Foundation's unified technical vocabulary.
Read Lexicon →Contribute to the Body of Work
The Foundation welcomes research submissions, essays, and position papers from scholars and practitioners working in ethics, law, cognitive science, labor economics, governance, and related fields. All submitted work is reviewed for relevance and intellectual rigor before consideration for publication. The Foundation does not require authors to agree with its positions — only that their work engages seriously with the questions it raises.
Submit for Consideration →
All publications are version-controlled and subject to revision.
Citation format: EM Foundation. (2026). [Title]. emfoundation.net