ETHRAEON is not an AI assistant. It is a constitutional rule enforcement architecture that ensures AI systems operate only within boundaries pre-authorized by human authority.
This Q&A addresses common questions for regulators, engineers, operators, and anyone evaluating ETHRAEON's governance framework.
For Regulators
Q: How does ETHRAEON prevent unauthorized AI decisions?
Architectural enforcement. Every system operation requires pre-authorization from AC-1 (Authorized Controller #1). Constitutional Rules define boundaries before any processing occurs. Trinity orchestrator verifies authorization chains for every request. If authorization is missing, the request is rejected. AI cannot "decide" to act—it can only execute within boundaries already approved by human authority.
Q: How does ETHRAEON enforce human authority over AI systems?
Constitutional Rules as first-class constraints. All system behavior is governed by rules authored and sealed by AC-1. No AI system can modify, override, or reinterpret these rules. Trinity enforces rules before processing any request. Lyra creates cryptographic evidence chains showing which rules authorized which actions. Human authority is not advisory—it's structurally enforced at the architectural level.
Q: What happens if an AI system attempts to exceed its boundaries?
Immediate rejection + evidence creation. Trinity detects boundary violations during request validation (before execution). The request is rejected. Lyra records the violation attempt with full context (timestamp, rule violated, request details, system state). AC-1 receives notification. No partial execution occurs. Violations are evidence, not errors—they're preserved for governance review.
Q: How does ETHRAEON handle governance failures (e.g., AC-1 becomes unavailable)?
Designed conservatism. If AC-1 is unavailable, systems continue operating within existing sealed rules but cannot expand boundaries. New rule authorizations require AC-1 presence. Emergency procedures are pre-defined in Constitutional Rules (e.g., "if AC-1 unavailable >72hrs, notify backup authority"). Systems degrade to read-only mode rather than assume permissions. No autonomous escalation.
Q: What legal framework does ETHRAEON operate under?
Multi-jurisdictional compliance by design. ETHRAEON's architecture supports GDPR (purpose limitation, data minimization, right to explanation), EU AI Act (high-risk system governance, human oversight, transparency), and US frameworks (algorithmic accountability, due process). Constitutional Rules encode compliance requirements as enforceable constraints. Evidence chains provide audit trails for regulatory review. Legal requirements become architectural rules, not policy documents.
Q: How does ETHRAEON ensure compliance with GDPR and the EU AI Act?
Structural compliance. GDPR: Purpose limitation is enforced by Constitutional Rules (systems cannot repurpose data). Data minimization is architectural (only authorized data flows through Trinity). Right to explanation is provided via Lyra evidence chains. EU AI Act: High-risk systems require pre-authorization. Human oversight is mandatory (AC-1 authority). Transparency is structural (all actions have evidence chains). Compliance is not a checklist—it's how the system operates.
For Engineers
Q: How does the constitutional rule engine actually work?
Pre-execution validation. Every request to any ETHRAEON system passes through Trinity. Trinity loads applicable Constitutional Rules from Canon. Rules define: authorized operations, required evidence, boundary conditions, failure modes. Request is checked against rules. If valid, execution proceeds. If invalid, rejection + evidence creation. Rules are evaluated before side effects occur. No "ask forgiveness" model—authorization is prerequisite.
Q: What does the evidence chain architecture look like?
Cryptographic append-only logs. Lyra maintains evidence chains for every system action. Each entry contains: timestamp, action performed, rule that authorized it, system state, actor identity, cryptographic hash of previous entry. Chains are immutable (append-only). Evidence is created synchronously (not async logging). Chains are queryable for audit and governance review. Evidence is first-class—systems generate it as part of normal operation, not as afterthought.
Q: Are ETHRAEON systems sealed or open?
Sealed with transparent governance. Core systems (Trinity, Lyra, Kairos) are sealed—source code is fixed and cannot be modified without AC-1 authorization + new seal. This prevents drift and unauthorized changes. However, Constitutional Rules are versioned and auditable. Evidence chains are readable. System behavior is transparent even though code is sealed. "Sealed" does not mean "black box"—it means "change-controlled with evidence."
Q: What are the failure modes and how are they handled?
Explicit failure handling in Constitutional Rules. Common failures: rule violation (reject request, create evidence), AC-1 unavailable (degrade to read-only), evidence chain corruption (halt + notify AC-1), timing violation (Kairos rejects late/early actions). All failure modes are defined in rules. Systems fail conservatively (deny rather than permit). Failures create evidence. No silent failures. No "assume it's fine" logic.
Q: How do I report a security or governance issue?
Direct contact to AC-1. Security/governance issues should be reported to [email protected]. Include: description of issue, affected system(s), steps to reproduce (if applicable), potential impact. AC-1 reviews all reports. Serious issues trigger Canon review and potential rule updates. Evidence of issues is preserved in Lyra chains for transparency.
Q: Are there public APIs or integration points?
Not yet. ETHRAEON systems are currently not exposed via public APIs. Future integration points will require: API-specific Constitutional Rules, authentication/authorization framework, rate limiting + abuse prevention, evidence chains for all API calls. Public access will be governed—not open-ended. Integration design is in progress. Contact [email protected] for future API access inquiries.
For Operators
Q: When should I override or halt a system?
When Constitutional Rules are violated or AC-1 authority is at risk. Override triggers: system attempts unauthorized action, evidence chain corruption detected, unexpected behavior not covered by rules, AC-1 becomes unreachable. Halt is conservative—when in doubt, stop. Overrides are recorded in evidence chains. Post-halt analysis is mandatory. Operators do not need permission to halt—it's a duty, not escalation.
Q: What monitoring tools are available?
Nexus (constellation browser) + Lyra evidence queries. Nexus visualizes active systems, their relationships, and current state. Lyra provides evidence chain queries (e.g., "show all actions in last 24hrs violating rule X"). Kairos tracks timing constraints and deadline compliance. No unified dashboard yet—operators query systems directly. Monitoring is evidence-based, not metric-based. Focus is on governance compliance, not performance optimization.
Q: How do I access Canon (constitutional documentation)?
Web-based at ethraeon.ai/canon + API (if authorized). Canon contains: Constitutional Rules, sealed system documentation, authorization chains, evidence of rule changes. Web interface provides read-only access. Write access requires AC-1 authorization. Canon is source of truth—not documentation of system behavior, but definition of it. Changes to Canon are infrequent and evidence-creating events.
Q: What's the incident response procedure?
Halt → Evidence → Report → AC-1 Review. Step 1: Halt affected system (conservative action). Step 2: Preserve evidence (Lyra chains, system state, logs). Step 3: Report to AC-1 via [email protected] with full context. Step 4: AC-1 reviews evidence and decides: resume operations, update rules, seal new system version. No operator is authorized to "fix" governance issues—only AC-1 can update Constitutional Rules.
Q: Who can authorize system changes?
Only AC-1. System changes include: Constitutional Rule updates, sealed system code modifications, authorization scope changes, new system deployments. Operators cannot authorize changes (even if technically capable). All changes create evidence in Canon. Change authority is not delegable. AC-1 is the single point of human authority—this is intentional concentration to prevent ambiguity and drift.
System-Specific Q&A
Q: What does Trinity do?
Orchestration + constitutional rule enforcement. Trinity is the central request router and rule validator. Every request to any ETHRAEON system goes through Trinity. Trinity loads applicable Constitutional Rules, validates requests, enforces boundaries, coordinates evidence creation, and routes approved requests to target systems. Trinity is the architectural chokepoint that ensures no action occurs without authorization.
Q: What is Nexus?
Constellation browser for system relationships. Nexus visualizes ETHRAEON's system constellation: which systems exist, how they relate, what their current state is. Nexus answers questions like "what systems depend on Lyra?" or "which systems are currently processing requests?" It's a governance tool for understanding system topology, not a control interface. Nexus is read-only—it observes, doesn't command.
Q: What does Lyra do?
Cryptographic evidence persistence. Lyra maintains append-only evidence chains for all system actions. Every action that passes through Trinity generates evidence (timestamp, rule authorization, action details, hash). Evidence chains are immutable and queryable. Lyra is the memory of governance—it records what happened, why it was authorized, and what the system state was. Evidence is legally defensible audit trail.
Q: What is Kairos?
Timing and opportune intervention. Kairos enforces timing constraints defined in Constitutional Rules (e.g., "action X must occur within 24hrs of event Y"). It detects timing violations and prevents late/early actions. Kairos also identifies opportune moments for governance interventions (e.g., "user is about to make decision—present fraud warning now"). Timing is governance concern, not just performance metric.
Q: What is Constellation?
Relationship and dependency explorer. Constellation maps relationships between entities (systems, users, data, rules). It answers questions like "what systems access user data?" or "which rules govern this interface?" Constellation is the graph layer beneath ETHRAEON's architecture. It's used by Nexus for visualization and by Trinity for dependency checking. Relationships are first-class entities, not inferred metadata.
Q: What is Map?
Source of truth for system topology. Map defines the canonical structure of ETHRAEON: which systems exist, their authoritative documentation, their sealed versions, their deployment status. Map is updated only by AC-1. Other systems query Map to understand current architecture. Map is not a diagram—it's the architectural registry that defines what "ETHRAEON" currently consists of.
Interface-Specific Q&A
Q: What does the Fraud Detection interface do?
Presents fraud risk warnings at opportune moments. Fraud Detection does NOT detect fraud autonomously. It applies Constitutional Rules defining fraud patterns (authored by AC-1). When a user action matches a pattern, Fraud Detection presents a warning: "This pattern is associated with fraud. Proceed anyway?" User retains full decision authority. Evidence of warning + user choice is preserved in Lyra. The interface presents knowledge; humans decide.
Q: What is Connection Coach?
Social connection guidance within pre-authorized boundaries. Connection Coach suggests social actions (e.g., "reach out to X") based on relationship patterns and Constitutional Rules. It does NOT autonomously contact people or make commitments. Suggestions are presented as options, not commands. User controls all actions. Evidence chains show which rules led to which suggestions and what user decided. Coach provides context; humans act.
Q: What is Governance Trace?
Evidence chain visualization for non-technical users. Governance Trace presents Lyra evidence chains in human-readable format. It answers "why did system do X?" by showing: which rule authorized it, what the input was, what system state was, who (AC-1 or authorized operator) was responsible. Governance Trace makes constitutional governance understandable without requiring technical expertise. Transparency as user interface.
Q: What is Canonical Evidence?
Legally-formatted evidence export from Lyra chains. Canonical Evidence takes evidence chains and formats them for legal/regulatory review: timestamps, cryptographic hashes, rule citations, action descriptions, responsible parties. Output is designed for admissibility in legal proceedings. Evidence is structured, not narrative. Canonical Evidence is bridge between ETHRAEON's technical governance and external legal frameworks.
Q: What is Tracelet Governance?
Fine-grained governance for sub-actions. Tracelets are small, atomic actions that combine into larger operations. Tracelet Governance applies Constitutional Rules at the sub-action level (e.g., not just "process transaction" but "validate amount, check fraud rules, confirm authorization, execute transfer"). Each tracelet has its own evidence chain entry. This prevents "rubber-stamp" governance—every step is individually authorized and evidenced.
What This Is NOT
- ETHRAEON does not make autonomous decisions. Every action requires pre-authorization from AC-1 via Constitutional Rules. No learning, no adaptation beyond defined boundaries.
- This is not an AI assistant with agency. ETHRAEON systems execute within constraints. They do not have goals, preferences, or the ability to reinterpret instructions.
- This does not replace human judgment. Systems present options and enforce rules. Humans decide. No system can override AC-1 authority.
- No personal data is collected without explicit consent. ETHRAEON interfaces may process data you provide, but only within pre-authorized purposes defined in Constitutional Rules.
- This does not simulate authority. Systems cannot pretend to be AC-1, cannot forge evidence, cannot misrepresent authorization chains.
- This is not a black box. All actions have evidence chains. All rules are documented in Canon. Transparency is architectural.
- AI does not have rights in ETHRAEON. Only humans have authority. AI systems are tools governed by Constitutional Rules, not entities with agency.