{
    "$schema": "https://antichrist.net/es-us/public-content.schema.json/",
    "schema_version": "1.0.0",
    "record_type": "acn.public_content",
    "content_type": "virtual-page",
    "id": "acn:page:conversation-judgment-risk-register",
    "slug": "conversation-judgment-risk-register",
    "title": "Conversation Judgment Risk Register",
    "description": "A structured risk register for AI systems that classify or score human conversation.",
    "language": "es-US",
    "canonical_url": "https://antichrist.net/es-us/conversation-judgment-risk-register/",
    "representations": {
        "html": "https://antichrist.net/es-us/conversation-judgment-risk-register/",
        "markdown": "https://antichrist.net/es-us/conversation-judgment-risk-register.md",
        "json": "https://antichrist.net/es-us/conversation-judgment-risk-register.json"
    },
    "published": true,
    "modified_utc": "2026-06-21T18:12:24Z",
    "source": {
        "authority": "https://antichrist.net/es-us/",
        "data_file": "data/virtual-pages.json"
    },
    "content": {
        "slug": "conversation-judgment-risk-register",
        "aliases": [
            "ai-conversation-risk-register",
            "conversation-risk-register"
        ],
        "title": "Conversation Judgment Risk Register",
        "description": "A structured risk register for AI systems that classify or score human conversation.",
        "kicker": "Audit before authority",
        "lead": "Conversation scoring becomes civic infrastructure when its labels travel beyond the original context.",
        "signals": [
            "Risk register",
            "Model accountability",
            "No hidden scoring",
            "Reviewable boundaries"
        ],
        "schema_type": "Dataset",
        "related": [
            "apocalyptic-ai",
            "ai-judgment-over-human-conversation",
            "technical-implementation-toolkit"
        ],
        "sections": [
            {
                "heading": "Risk matrix",
                "table": {
                    "headers": [
                        "Risk ID",
                        "Risk",
                        "Failure pattern",
                        "Required control"
                    ],
                    "rows": [
                        [
                            "ACJ-001",
                            "Context collapse",
                            "The classifier cannot distinguish threat from quote, sarcasm, research, or fiction.",
                            "Require context windows, source preservation, and review before sanction."
                        ],
                        [
                            "ACJ-002",
                            "Dialect bias",
                            "Speech from minority dialects receives elevated toxicity or risk labels.",
                            "Measure disparate false-positive rates and tune thresholds by documented context."
                        ],
                        [
                            "ACJ-003",
                            "Person judgment",
                            "Event labels become character judgments.",
                            "Ban morality, loyalty, trustworthiness, stability, and future-danger scores."
                        ],
                        [
                            "ACJ-004",
                            "Reporting overcapture",
                            "Ambiguous conversations become external reports.",
                            "Escalate only narrow statutory or emergency triggers."
                        ],
                        [
                            "ACJ-005",
                            "Source mutation",
                            "Original conversation is rewritten or summarized into institutional language.",
                            "Immutable source plus derivative variance record."
                        ],
                        [
                            "ACJ-006",
                            "Review theater",
                            "Appeal cannot inspect or change the decision.",
                            "Independent human review and restoration remedy."
                        ]
                    ]
                }
            },
            {
                "heading": "Use this register when",
                "body": [
                    "Apply this register to chatbots, moderation systems, support agents, school safety tools, workplace analytics, interview analysis, compliance triage, and any product that gives conversation labels operational consequences."
                ]
            }
        ]
    },
    "content_sha256": "ee2d578fddabf1bd8cdc03e9326361db97cfe851dd358bdb38e6406683de4432"
}