Network Working Group                                         Y. Wang
Internet-Draft                                         HJS Foundation
Intended status: Informational                                   Ltd.
Expires: 20 September 2026                                21 March 2026

                     Co-Evolve Binding Protocol-00 (CEP)
                  draft-wang-cep-00

Abstract

   This document specifies the Co-Evolve Binding Protocol (CEP), a
   lightweight, minimal-structured protocol designed to bind human
   cognitive evolution with artificial intelligence (AI) evolution.
   CEP works in conjunction with the Judgment Event Protocol (JEP) and
   Human Judgment Structure (HJS) as the default collaborative system,
   while supporting adaptation to other compliant judgment and
   accountability systems to enhance interoperability and practicality.
   Leveraging JEP's strengths in standardized judgment recording and
   lightweight verifiability, HJS's advantages in structured human
   responsibility management and end-to-end accountability traceability,
   and the flexibility to adapt to other systems, CEP integrates core
   advantages to form a complementary three-layer human-AI symbiosis
   infrastructure. The core goal of CEP is to prevent unilateral AI
   evolution, maintain human sovereignty in the co-evolution process of
   humans and AI, and lay the foundation for trustworthy human-AI
   symbiosis.

Status of This Memo

   This Internet-Draft is submitted in full conformance with the
   provisions of BCP 78 and BCP 79.

   Internet-Drafts are working documents of the Internet Engineering
   Task Force (IETF). Note that other groups may also distribute
   working documents as Internet-Drafts. The list of current Internet-
   Drafts is at https://datatracker.ietf.org/drafts/current/.

   Internet-Drafts are draft documents valid for a maximum of six months
   and may be updated, replaced, or obsoleted by other documents at any
   time. It is inappropriate to use Internet-Drafts as reference
   material or to cite them other than as "work in progress."

   This Internet-Draft will expire on 20 September 2026.

Copyright Notice

   Copyright (c) 2026 IETF Trust and the persons identified as the
   document authors. All rights reserved.

   This document is subject to BCP 78 and the IETF Trust's Legal
   Provisions Relating to IETF Documents
   (https://trustee.ietf.org/license-info) in effect on the date of
   publication of this document. Please review these documents
   carefully, as they describe your rights and restrictions with respect
   to this document.

Table of Contents

   1. Introduction .................................................... 3
      1.1. Background ................................................ 3
      1.2. Objectives ................................................ 4
      1.3. Scope ..................................................... 5
   2. Terminology ..................................................... 6
   3. Protocol Overview ............................................... 7
      3.1. Architecture Positioning .................................. 7
      3.2. Core Design Principles .................................... 8
   4. Protocol Specification ......................................... 10
      4.1. Core Fields (Mandatory) .................................. 10
         4.1.1. HumanAnchor ......................................... 11
         4.1.2. ShiftClass .......................................... 11
         4.1.3. BoundState .......................................... 12
      4.2. Optional Extended Fields ................................. 13
      4.3. Protocol Frame Format .................................... 14
      4.4. Protocol Interaction Process ............................. 15
      4.5. Adaptation Specifications for Other Compliant Systems .... 16
   5. Security Considerations ....................................... 18
   6. Privacy Considerations ........................................ 18
   7. IANA Considerations ........................................... 19
   8. References .................................................... 19
      8.1. Normative References ..................................... 19
      8.2. Informative References ................................... 20
   Author's Address .................................................. 20

1.  Introduction

1.1.  Background

   With the rapid development of AI technology, especially the emergence
   of large language models and autonomous agents, AI is increasingly
   moving towards self-evolution, self-optimization, and closed-loop
   reasoning. This trend brings the risk of AI divergence from human
   cognitive logic, value systems, and responsibility boundaries—
   leading to a "civilization bifurcation" between human civilization
   and silicon-based intelligence.

   To address this risk, the Judgment Event Protocol (JEP)
   [I-D.wang-jep-judgment-event-protocol] provides a minimal, verifiable
   format for logging judgment-related actions, ensuring standardized
   human participation in key AI decision-making processes with its
   lightweight design and replay protection mechanism. The Human
   Judgment Structure (HJS) [I-D.wang-hjs-judgment-structure], as an
   infrastructure-grade accountability layer built on JEP, complements
   JEP's advantages by providing a structured framework for human
   responsibility management, cryptographic accountability mechanisms,
   and a three-tier privacy architecture, ensuring traceability of
   causality and accountability in AI operations. However, there is a
   lack of a dedicated protocol to bind human and AI evolution, leading
   to potential unilateral AI evolution and loss of human sovereignty in
   the evolution process—CEP is designed to fill this gap by integrating
   JEP and HJS's core strengths, while supporting adaptation to other
   compliant judgment and accountability systems to meet diverse
   practical deployment needs.

1.2.  Objectives

   CEP is designed to achieve the following core objectives, aligning
   with JEP, HJS and other compliant systems, and integrating their
   strengths to form a three-layer human-AI symbiosis infrastructure
   with strong adaptability:

   1.  Bind AI evolution to human cognitive evolution, preventing
       unilateral AI evolution and civilization bifurcation, leveraging
       JEP's standardized judgment records, HJS's structured
       accountability, and the adaptability to other systems to ensure
       traceable and controllable co-evolution.

   2.  Provide minimal-structured, extensible fields to cover all AI
       evolution scenarios (e.g., fine-tuning, emergence, drift, loop
       formation), inheriting JEP's minimal design concept, HJS's
       extensible architecture, and supporting flexible adaptation to
       other systems without modifying the core structure.

   3.  Ensure seamless compatibility with JEP, HJS, and other compliant
       judgment/accountability systems, integrating their verification
       mechanisms and accountability traceability capabilities to enable
       end-to-end human-AI co-evolution, decision-making, and
       accountability.

   4.  Be lightweight, easy to implement, and compatible with existing
       AI systems, network protocols, and W3C standards, adhering to
       JEP's lightweight engineering design, HJS's infrastructure-grade
       compatibility requirements, and the practical deployment needs of
       other systems.

1.3.  Scope

   CEP applies to all AI systems, large models, and intelligent agents
   that involve self-evolution, self-optimization, or capability
   emergence. It focuses on the binding of human and AI evolution
   processes, without involving AI model training, algorithm design, or
   hardware implementation. Drawing on JEP's non-intrusive design,
   HJS's middleware-style deployment, and the flexible adaptation
   capability to other systems, CEP works as a middleware layer,
   inserting into the evolution process of AI systems to ensure human
   oversight and symbiotic evolution, while maintaining compatibility
   with JEP's judgment recording, HJS's accountability traceability, and
   the core functions of other compliant systems.

2.  Terminology

   The following terms are used throughout this document:

   CEP
         Co-Evolve Binding Protocol, the protocol specified in this
         document, which integrates the strengths of JEP, HJS and other
         compliant systems to bind human and AI evolution.

   JEP
         Judgment Event Protocol [I-D.wang-jep-judgment-event-protocol],
         a minimal, verifiable format for logging judgment-related
         actions in distributed systems, which defines four immutable
         event verbs (J, D, T, V), a signed JSON event structure with
         replay protection, and standardized signature verification
         rules. Its core strengths lie in lightweight design,
         standardized judgment recording, and reliable verification,
         serving as the default judgment system for CEP and providing
         structured human judgment records and a verification interface.

   HJS
         Human Judgment Structure [I-D.wang-hjs-judgment-structure], an
         infrastructure-grade accountability layer for AI agents built
         on the Judgment Event Protocol (JEP), which inherits JEP's four
         core primitives (Judge, Delegate, Terminate, Verify) and
         complements JEP's advantages with a structured human
         responsibility management framework, a cryptographic mechanism
         for cross-platform responsibility verification, a minimal
         receipt structure with responsibility chaining, and a three-
         tier privacy architecture for accountability governance. It
         serves as the default accountability system for CEP, fully
         compatible with CEP's human-AI co-evolution binding logic,
         providing end-to-end accountability traceability.

   Compliant Judgment/Accountability Systems
         Other judgment or accountability systems that meet international
         standards (e.g., RFC, W3C) and core requirements of human-AI
         symbiosis (e.g., human sovereignty, traceability,
         accountability), which CEP can adapt to on the premise of not
         changing its core structure.

   HumanAnchor
         A core field in CEP, used to bind AI evolution to a specific
         human judgment or value anchor (associated with the judgment ID
         of JEP or other compliant judgment systems), leveraging the
         standardized judgment recording of JEP/other systems and the
         responsibility binding of HJS/other compliant accountability
         systems to ensure traceability of the binding relationship.

   ShiftClass
         A core field in CEP, used to classify the type of AI evolution
         (e.g., fine-tuning, emergence, drift), enabling targeted
         supervision and aligning with the structured accountability
         management requirements of HJS and other compliant
         accountability systems.

   BoundState
         A core field in CEP, used to mark whether AI evolution is bound
         to human cognitive evolution (enabling symbiotic evolution),
         serving as a "switch" for co-evolution and integrating the
         responsibility closure mechanism of HJS and other compliant
         accountability systems to ensure that unbound unilateral
         evolution is rejected.

   Co-Evolution
         The process in which AI and humans evolve synchronously, with
         AI's evolution amplifying human cognitive capabilities and
         human's evolution constraining AI's boundary, supported by the
         judgment participation of JEP/other compliant systems and the
         accountability traceability of HJS/other compliant systems.

   Unilateral Evolution
         The process in which AI evolves independently without human
         oversight, value alignment, or responsibility binding, which
         CEP prevents by integrating the joint constraints of JEP, HJS,
         and other compliant systems.

3.  Protocol Overview

3.1.  Architecture Positioning

   CEP is the third layer of the human-AI symbiosis infrastructure,
   working with JEP, HJS, and other compliant judgment/accountability
   systems, and integrating their core strengths to form a three-layer
   closed-loop system with strong adaptability. Each layer complements
   the others, with JEP/other compliant judgment systems providing
   standardized judgment support, HJS/other compliant accountability
   systems offering structured accountability protection, and CEP
   realizing co-evolution binding, forming a complete human-AI symbiosis
   system:

   Bottom Layer: HJS and other compliant accountability systems. HJS, as
      the default infrastructure-grade accountability layer for AI
      agents built on JEP, inherits JEP's lightweight and verifiable
      advantages and complements them with a structured human
      responsibility management framework. Other compliant
      accountability systems that meet core requirements can also be
      adapted, providing a cryptographic mechanism to manage and verify
      responsibility for AI agent judgment behaviors across
      heterogeneous platforms, defining core primitives compatible with
      CEP, with a minimal receipt structure and privacy architecture for
      accountability governance. They enable traceability of AI
      evolution events, complete the closed-loop of human-AI co-
      evolution accountability, and provide reliable accountability
      support for CEP.

   Middle Layer: JEP and other compliant judgment systems. JEP, with its
      core strengths of lightweight design, standardized judgment
      recording, and reliable verification, ensures human participation
      in key decision-making, serves as the default judgment system for
      CEP, and provides structured human judgment records and a
      verification interface. Other compliant judgment systems that meet
      core requirements can also be adapted, providing standardized
      human judgment records compatible with CEP's HumanAnchor field,
      laying the foundation for CEP's co-evolution binding.

   Top Layer: CEP (Co-Evolve Binding Protocol). Integrates the
      standardized judgment verification of JEP/other compliant judgment
      systems and the structured accountability traceability advantages
      of HJS/other compliant accountability systems, binds human and AI
      evolution, locks the "non-divergence" of human and AI
      civilizations, and realizes the organic connection between
      judgment participation and accountability closure of various
      compliant systems.

3.2.  Core Design Principles

   CEP adheres to the following design principles, consistent with
   IETF's preference for lightweight, engineering-oriented, and
   standardized protocols, while fully integrating the core strengths of
   JEP, HJS, and other compliant systems:

   Minimal Structuring: Adopt a minimal-structured design consistent
      with JEP's lightweight concept, HJS's minimal receipt structure,
      and the practical needs of other compliant systems, using only 3
      core mandatory fields to avoid redundant design, reduce the cost
      of protocol implementation and maintenance, and align with the
      cost optimization requirements of all compliant systems, forming a
      complementary relationship with their cryptographic erasure and
      privacy architecture to achieve cost structure optimization.

   Interoperability: Be compatible with existing AI systems, network
      protocols, JEP, HJS, and other compliant judgment/accountability
      systems, leveraging their standardized formats and cross-platform
      compatibility to enable seamless integration and global
      interoperability.

   Engineering-Oriented: Use mature IETF terminology (e.g., Binding,
      State) to avoid abstract or biological terms, following the
      engineering-oriented design concept of JEP, HJS, and other
      compliant systems, ensuring readability and implementability by
      engineers.

   Extensibility and Adaptability: Support optional extended fields to
      adapt to future AI evolution scenarios and the characteristics of
      other compliant systems without changing the core structure,
      inheriting the extensible architecture of HJS, the flexible
      verification mechanism of JEP, and the adaptation needs of other
      systems.

   Human Sovereignty: Ensure that human beings have the final right to
      control AI evolution, and AI evolution cannot be separated from
      human oversight, integrating the human judgment participation of
      JEP/other compliant judgment systems and the human responsibility
      management advantages of HJS/other compliant accountability
      systems to safeguard human sovereignty in co-evolution.

4.  Protocol Specification

4.1.  Core Fields (Mandatory)

   CEP defines 3 mandatory core atomic fields, which are minimal,
   structured, and compatible with JEP, HJS, and other compliant
   judgment/accountability systems. All AI evolution events must carry
   these 3 fields to ensure symbiotic binding. The design of these
   fields fully integrates the standardized advantages of JEP/other
   compliant judgment systems and the accountability requirements of
   HJS/other compliant accountability systems, ensuring lightweight
   implementation while maintaining traceability and controllability.

4.1.1.  HumanAnchor

   Definition: The unique identifier of the human judgment or value
      anchor bound to AI evolution, directly associated with the
      judgment event ID (or equivalent core identifier) in JEP or other
      compliant judgment systems. This field integrates the standardized
      judgment recording advantage of JEP/other compliant judgment
      systems and the responsibility binding requirement of HJS/other
      compliant accountability systems, ensuring that AI evolution is
      anchored to human cognitive logic and value boundaries, and the
      validity of this identifier is verified through the standardized
      signature verification rules of JEP/other compliant judgment
      systems, while the binding relationship is traceable through the
      accountability system of HJS/other compliant accountability
      systems.

   Format: Consistent with the judgment event identifier format of JEP
      or other compliant judgment systems (for JEP, this is the UUIDv4
      nonce field or the multihash "what" field of JEP events; for other
      compliant systems, as defined in their core specifications),
      ensuring interoperability between CEP and JEP/other compliant
      judgment systems, and compatible with the accountability log ID
      format of HJS/other compliant accountability systems to facilitate
      traceability.

   Constraints: Cannot be empty; must correspond to a valid judgment
      event in JEP or other compliant judgment systems (verified via the
      signature verification steps of the corresponding system,
      including nonce uniqueness, timestamp validity, and signature
      correctness as defined in Section 2.4 of [I-D.wang-jep-judgment-
      event-protocol]), and the associated human responsibility
      information must be recorded in HJS or other compliant
      accountability systems to ensure accountability closure. If the
      associated judgment event is invalid or the responsibility record
      is missing, the AI evolution event is rejected.

4.1.2.  ShiftClass

   Definition: Classifies the type of AI evolution, used to identify the
      nature of AI evolution and implement targeted supervision. This
      field is designed in line with the structured accountability
      management advantage of HJS and other compliant accountability
      systems, enabling them to perform classified accountability based
      on evolution types, while inheriting the standardized enumeration
      design of JEP and other compliant systems to ensure consistency.

   Enumeration Values (fixed, non-extensible to ensure standardization,
      consistent with the standardized design concept of JEP, HJS, and
      other compliant systems):

      FINE_TUNE: AI model parameter fine-tuning, incremental evolution.

      EMERGENCE: Sudden emergence of new capabilities that were not
         explicitly trained.

      DRIFT: Deviation of AI's values, preferences, or reasoning logic
         from the original human alignment.

      LEAP: Qualitative leap in AI capabilities (e.g., from narrow AI to
         general AI in a specific field).

      LOOP: AI forms a closed-loop reasoning chain that is independent
         of human oversight.

   Constraints: Must be one of the above enumeration values; no custom
      values are allowed, ensuring standardization consistent with JEP,
      HJS, and other compliant systems, and facilitating unified
      supervision and accountability.

4.1.3.  BoundState

   Definition: Marks whether AI evolution is bound to human cognitive
      evolution, serving as a "switch" for symbiotic evolution. This
      field integrates the responsibility closure mechanism of HJS/other
      compliant accountability systems and the human judgment
      verification requirement of JEP/other compliant judgment systems,
      ensuring that only AI evolution bound to human cognitive evolution
      can take effect, preventing unilateral evolution.

   Values:

      True: AI evolution is bound to human cognitive evolution
         (synchronous co-evolution), and the evolution event is allowed
         to take effect. The binding relationship must be verified by
         JEP/other compliant judgment systems and recorded in HJS/other
         compliant accountability systems to ensure accountability.

      False: AI evolution is not bound to human cognitive evolution
         (unilateral evolution), and the evolution event is rejected, in
         line with the human participation requirement of JEP/other
         compliant judgment systems and the responsibility management
         principle of HJS/other compliant accountability systems.

   Optional Note: A short text description (≤ 100 characters) explaining
      the reason for BoundState (e.g., "Bound to JEP judgment ID: 12345,
      ensuring value alignment" or "Bound to [Other Compliant System]
      judgment ID: 67890, ensuring traceability"). This field is
      optional but recommended for accountability and traceability,
      consistent with the traceability requirement of HJS/other
      compliant accountability systems and the record completeness
      principle of JEP/other compliant judgment systems.

   Constraints: Must be a Boolean value; the evolution event is rejected
      if BoundState is False. If BoundState is True, the associated
      HumanAnchor must be valid (verified by JEP/other compliant
      judgment systems) and the responsibility information must be
      recorded in HJS/other compliant accountability systems, ensuring
      the closed-loop of binding and accountability.

4.2.  Optional Extended Fields

   To adapt to specific scenarios (e.g., compliance, traceability), CEP
   supports optional extended fields. These fields do not change the
   core structure of CEP and are compatible with the mandatory fields,
   while supporting adaptation to JEP, HJS, and other compliant systems.
   The design of extended fields fully considers the complementary
   advantages of all compliant systems, further enhancing the
   traceability and accountability of CEP.

   EchoLogID (String): Unique identifier of the evolution log,
      associated with the accountability log ID of HJS or other
      compliant accountability systems, which can be directly
      synchronized to their AI agent judgment accountability traceability
      system.

   EvolveTime (Timestamp): The time when the AI evolution event occurs
      (format: ISO 8601), consistent with the timestamp format of JEP/
      other compliant judgment systems.

   OperatorID (String): Unique identifier of the human operator,
      associated with the responsible person ID of HJS/other compliant
      accountability systems and the judgment executor ID of JEP/other
      compliant judgment systems.

   SystemAdaptor (String, Optional): Unique identifier of the compliant
      judgment/accountability system adapted by CEP (e.g., "JEP+HJS",
      "[Other Judgment System]+[Other Accountability System]").

4.3.  Protocol Frame Format

   CEP uses a lightweight frame format, compatible with JSON and
   Protocol Buffers (Protobuf) to adapt to different network
   environments and AI systems. The core frame format inherits the
   lightweight JSON/Protobuf compatibility of JEP/other compliant
   judgment systems and the structured data organization advantage of
   HJS/other compliant accountability systems, ensuring readability,
   high performance, and compatibility with all compliant systems. The
   core frame format is as follows (JSON example, including adaptation
   to JEP+HJS and other systems):

   {
     "CEP": {
       "version": "0.1",
       "systemAdaptor": "JEP+HJS",
       "coreFields": {
         "HumanAnchor": "f47ac10b-58cc-4372-a567-0e02b2c3d479",
         "ShiftClass": "EMERGENCE",
         "BoundState": {
           "value": true,
           "note": "Bound to JEP nonce: f47ac10b-58cc-4372-a567-0e02b2c3d479"
         }
       },
       "extendedFields": {
         "EchoLogID": "log:789e4567-e89b-12d3-a456-426614174000",
         "EvolveTime": "2026-03-21T10:00:00Z",
         "OperatorID": "did:example:operator-98765"
       }
     }
   }

   Note: For JEP association, the HumanAnchor field SHOULD use the JEP
   event's nonce (UUIDv4) as the primary identifier, as defined in
   Section 2.2 of [I-D.wang-jep-judgment-event-protocol]. The multihash
   "what" field MAY be used as an alternative reference when appropriate.

4.4.  Protocol Interaction Process

   CEP's interaction process is simple and lightweight, involving three
   core roles: AI System, compliant judgment system (JEP or other), and
   compliant accountability system (HJS or other). The process fully
   integrates the verification advantage of compliant judgment systems
   and the accountability traceability advantage of compliant
   accountability systems, ensuring the validity of co-evolution binding
   and the completeness of accountability closure. The process is as
   follows:

   1.  The AI System detects an evolution event (e.g., capability
       emergence, parameter fine-tuning, logic drift), determines the
       compliant judgment/accountability system to be used (defaults to
       JEP+HJS), generates CEP frame with mandatory core fields
       (HumanAnchor, ShiftClass, BoundState) and optional extended
       fields (if needed), and associates the HumanAnchor with the valid
       judgment ID (e.g., JEP nonce) from the selected compliant
       judgment system.

   2.  The AI System sends the generated CEP frame to the selected
       compliant judgment system (JEP or other) to verify the validity
       of the HumanAnchor field. The judgment system performs
       verification in accordance with its standardized signature
       verification rules (for JEP, as defined in Section 2.4 of
       [I-D.wang-jep-judgment-event-protocol], including nonce
       uniqueness, timestamp validity, and signature correctness) to
       confirm that the associated judgment event is valid.

   3.  If the verification fails (e.g., invalid HumanAnchor,
       unassociated judgment event), the compliant judgment system
       returns a verification failure message to the AI System, and the
       AI System rejects the AI evolution event, terminating the
       interaction process.

   4.  If the verification succeeds, the compliant judgment system
       returns a verification success message, and the AI System
       executes the AI evolution event (only when BoundState is True; if
       BoundState is False, the evolution event is directly rejected
       regardless of verification result).

   5.  After the evolution event is executed (or rejected), the AI
       System synchronizes the CEP frame (including evolution result) to
       the selected compliant accountability system (HJS or other). The
       accountability system records the CEP evolution information,
       associates it with the corresponding judgment record and human
       responsibility information from the compliant judgment system,
       and completes the accountability traceability entry in accordance
       with its structured responsibility management framework and
       privacy architecture.

   6.  The compliant accountability system returns a synchronization
       success message to the AI System, and the CEP interaction process
       is completed.

4.5.  Adaptation Specifications for Other Compliant Systems

   To ensure the compatibility and reliability of CEP when adapting to
   other compliant judgment/accountability systems (excluding JEP and
   HJS), the following adaptation specifications must be met to avoid
   conflicts with CEP's core mechanism and ensure the achievement of
   human-AI symbiosis goals:

   1.  Core Requirement Compliance: The adapted judgment/accountability
       system must meet the core requirements of human-AI symbiosis,
       including but not limited to: ensuring human sovereignty in AI
       evolution, providing standardized human judgment records (for
       judgment systems) or structured accountability traceability (for
       accountability systems), supporting non-repudiation and tamper-
       proofing of data, and complying with international compliance
       standards (e.g., GDPR, EU AI Act).

   2.  Format Compatibility: The judgment ID format of the compliant
       judgment system must be compatible with CEP's HumanAnchor field
       (supporting UUIDv4, multihash, or other internationally
       standardized formats); the accountability log ID format of the
       compliant accountability system must be compatible with CEP's
       EchoLogID field, ensuring seamless association and data
       synchronization.

   3.  Verification Mechanism Compatibility: The compliant judgment
       system must provide a standardized signature verification
       mechanism (including nonce uniqueness, timestamp validity,
       signature correctness) consistent with CEP's verification
       requirements, ensuring that the validity of the HumanAnchor field
       can be verified without modifying CEP's core logic.

   4.  Data Synchronization Compatibility: The compliant accountability
       system must support the synchronization of CEP evolution records,
       be able to associate with the judgment records of the selected
       compliant judgment system, and complete the entry of
       accountability traceability, ensuring the closed-loop of
       "judgment-binding-accountability".

   5.  Non-Intrusive Adaptation: The adaptation process of other
       compliant systems must not modify CEP's core structure (3
       mandatory core fields) and core design principles, and must be
       implemented in a middleware-style deployment consistent with CEP,
       avoiding intrusion into CEP's internal logic or AI model
       structure.

5.  Security Considerations

   The security of CEP relies on the security of the underlying
   compliant judgment and accountability systems (JEP, HJS, or other
   adapted systems). Implementations MUST ensure:

   o  The integrity and authenticity of the HumanAnchor field are
      verified through the signature verification mechanisms of the
      associated compliant judgment system (for JEP, as defined in
      Section 2.4 of [I-D.wang-jep-judgment-event-protocol]).

   o  The BoundState field is enforced as a critical control point; AI
      evolution events with BoundState set to False MUST be rejected
      unconditionally.

   o  All CEP frames SHOULD be transmitted over secure channels (e.g.,
      TLS 1.3 or equivalent) to prevent man-in-the-middle attacks.

   o  Replay protection is provided by the underlying compliant judgment
      system (e.g., nonce uniqueness in JEP, as defined in Section 2.3
      of [I-D.wang-jep-judgment-event-protocol]) and MUST NOT be
      bypassed.

6.  Privacy Considerations

   CEP is designed to align with privacy requirements such as GDPR,
   including the right to erasure. Implementations SHOULD:

   o  Use the privacy architecture of HJS or other compliant
      accountability systems to manage human responsibility information,
      ensuring that personally identifiable information (PII) is
      protected.

   o  Support cryptographic erasure mechanisms where required, allowing
      the removal of human responsibility associations when legally
      mandated.

   o  Minimize the collection of human-identifiable data; the
      OperatorID field SHOULD be a pseudonymous identifier (e.g., DID)
      rather than directly identifying information.

7.  IANA Considerations

   This document has no IANA actions.

8.  References

8.1.  Normative References

   [RFC2119]  Bradner, S., "Key words for use in RFCs to Indicate
              Requirement Levels", BCP 14, RFC 2119,
              DOI 10.17487/RFC2119, March 1997,
              <https://www.rfc-editor.org/info/rfc2119>.

   [RFC8174]  Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC
              2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174,
              May 2017, <https://www.rfc-editor.org/info/rfc8174>.

   [RFC8259]  Bray, T., Ed., "The JavaScript Object Notation (JSON) Data
              Interchange Format", STD 90, RFC 8259,
              DOI 10.17487/RFC8259, December 2017,
              <https://www.rfc-editor.org/info/rfc8259>.

   [RFC8785]  Rundgren, A., Jordan, B., and S. Erdtman, "JSON
              Canonicalization Scheme (JCS)", RFC 8785,
              DOI 10.17487/RFC8785, June 2020,
              <https://www.rfc-editor.org/info/rfc8785>.

   [RFC9562]  Davis, K., Peabody, B., and P. Leach, "Universally Unique
              IDentifiers (UUIDs)", RFC 9562, DOI 10.17487/RFC9562,
              May 2024, <https://www.rfc-editor.org/info/rfc9562>.

   [ISO8601]  International Organization for Standardization, "Data
              elements and interchange formats -- Information
              interchange -- Representation of dates and times",
              ISO 8601-1:2019, February 2019.

8.2.  Informative References

   [I-D.wang-jep-judgment-event-protocol]
              Wang, Y., "Judgment Event Protocol (JEP)", Work in
              Progress, Internet-Draft, draft-wang-jep-judgment-event-
              protocol-01, 20 March 2026,
              <https://datatracker.ietf.org/doc/html/draft-wang-jep-
              judgment-event-protocol-01>.

   [I-D.wang-hjs-judgment-structure]
              Wang, Y., "Human Judgment Structure (HJS)", Work in
              Progress, Internet-Draft, draft-wang-hjs-judgment-
              structure-01, 20 March 2026,
              <https://datatracker.ietf.org/doc/html/draft-wang-hjs-
              judgment-structure-01>.

Author's Address

   Yuqiang Wang
   HUMAN JUDGMENT SYSTEMS FOUNDATION LTD.
   Email: signal@humanjudgment.org
   GitHub: https://github.com/hjs-spec