{
  "decision_id": "dvr_027DU3mGJkkOAraH",
  "reason_codes": [],
  "policy_epoch": "pe_sandbox_v1",
  "decided_at": "2026-07-10T00:40:37.396Z",
  "korero_sha256": "afa325c203b0d72d200a7c15efe954445f90bbbd8d4d71e1457d46941324d250",
  "tenant_did": "did:web:devtest.mysovereignty.digital",
  "record_type": "deliberation_record",
  "record_frozen": true,
  "context": {
    "tags": []
  },
  "korero": {
    "proposal": {
      "title": "Security & Resilience",
      "description": "Deliberation strand adversary"
    },
    "participants": [
      {
        "ref": "p1",
        "role_label": "member",
        "did": null
      },
      {
        "ref": "p2",
        "role_label": "facilitator",
        "did": null
      },
      {
        "ref": "p3",
        "role_label": "member",
        "did": null
      },
      {
        "ref": "p4",
        "role_label": "member",
        "did": null
      }
    ],
    "entries": [
      {
        "seq": 1,
        "at": "2026-07-10T00:37:59.500Z",
        "participant_ref": "p2",
        "kind": "statement",
        "body": "This strand deals with what can go wrong when someone else holds the switch. In June 2026 a single overseas order disrupted access to a major AI tool worldwide — that is the live demonstration of dependency risk [S-1]. Our first hyperscale cloud region, opened December 2024, is US-owned through an Irish subsidiary [S-15]: onshore racks, offshore control. And the human surface is wide open — 48% of workers have entered sensitive information into public AI tools; 56% used AI at work unsure whether they were allowed to [S-1]. We will be disciplined about what we do not know: there is no NZ-specific AI adversary-threat metric, and we will not invent one [G-4]. — Rewi Thompson-Hale (chair)",
        "refs": []
      },
      {
        "seq": 2,
        "at": "2026-07-10T00:37:59.520Z",
        "participant_ref": "p2",
        "kind": "statement",
        "body": "Let me put the dependency case first because it is the one with a live proof. June 2026: one overseas legal order, one major AI tool, disrupted worldwide access — no malice toward New Zealand required, we were simply downstream [S-1]. Now map our exposure: the first hyperscale region on our soil is owned by a US corporation through an Irish subsidiary [S-15]. Residency of the racks does not change who answers a foreign court order. In security terms this is not an adversary problem in the military sense — it is a jurisdiction and dependency problem, and the mitigation is architectural: know your dependencies, test where inference actually runs, and hold a fallback for services that cannot stop. The cross-party proposal's jurisdiction-of-inference procurement test [S-1] is, from a practitioner's seat, the single highest-value control on the table.",
        "refs": []
      },
      {
        "seq": 3,
        "at": "2026-07-10T00:37:59.536Z",
        "participant_ref": "p3",
        "kind": "statement",
        "body": "Rewi, your architecture is sound and I will support it — but the interference surface that worries me most needs no data centre at all. It runs through people and provenance. The numbers: 48% of workers have put sensitive information into public AI tools, and 56% have used AI at work without knowing whether they were allowed [S-1]. That is an exfiltration channel and a norms vacuum at national scale, today, at zero cost to any adversary. Add synthetic media in an election year and the cheapest attack on New Zealand is not turning anything off — it is polluting what people believe official information looks like. My claim: a provenance standard for official communications and a plain usage standard for workers are as much resilience instruments as your jurisdiction test, and they ship faster.",
        "refs": []
      },
      {
        "seq": 4,
        "at": "2026-07-10T00:37:59.551Z",
        "participant_ref": "p4",
        "kind": "statement",
        "body": "I will add the layer both of you route around: logistics. Every accelerator in every rack in this country arrived on a ship or a plane from a fabrication base concentrated in East Asia — New Zealand has no domestic fabrication and is wholly import-dependent [S-21]. From a supply-chain seat, resilience is not just who controls the software switch; it is who controls the replacement part, the refresh cycle, and the shipping lane. A dependency you have mapped is a risk; a dependency you have not mapped is a surprise. Right now no register exists of which critical services — health triage, border processing, energy dispatch — depend on which AI suppliers, models, and hardware. Before we argue about which control matters most, we should agree nobody can rank risks on dependencies nobody has listed.",
        "refs": []
      },
      {
        "seq": 5,
        "at": "2026-07-10T00:37:59.568Z",
        "participant_ref": "p2",
        "kind": "statement",
        "body": "GAP acknowledgement — G-4. The record should state what this strand refuses to do: there is no NZ-specific quantified AI adversary-threat metric, and we will not manufacture one to make our recommendations look load-bearing [G-4]. The operative, evidenced risk is jurisdictional and dependency-shaped — foreign corporate control of onshore infrastructure [S-15], demonstrated worldwide disruption from a single overseas order [S-1] — not a military threat model, and the draft should say so plainly. This discipline cuts both ways: it forbids threat inflation to justify spend, and it forbids complacency dressed as scepticism. Where we cannot quantify, we recommend instruments that create the missing visibility — which is exactly what Sefo's dependency register is.",
        "refs": []
      },
      {
        "seq": 6,
        "at": "2026-07-10T00:37:59.590Z",
        "participant_ref": "p3",
        "kind": "statement",
        "body": "Then let me sharpen my disagreement with the emerging ranking rather than let it pass politely. Rewi's jurisdiction test and Sefo's register are both right, and both are slow-burn instruments: procurement tests bind at the next contract renewal; registers populate over a year or more. An information-integrity incident does not wait for procurement cycles. The provenance standard — official communications carry verifiable provenance marks, agencies publish what synthetic content they will and will not produce — is cheap, visible to the public, and deployable in months. If this strand ranks it third because it feels soft next to infrastructure, we will have made the classic security-committee error: funding the threat that flatters our expertise over the one with the shortest fuse [S-1]. I want it ranked as a co-equal first-wave instrument, not an annex.",
        "refs": []
      },
      {
        "seq": 7,
        "at": "2026-07-10T00:37:59.612Z",
        "participant_ref": "p4",
        "kind": "statement",
        "body": "Ingrid, I will meet you halfway with a sequencing argument rather than a ranking one. The register is not slow if you scope wave one tightly: the top twenty critical-service dependencies can be mapped in a quarter — which supplier, which model, which jurisdiction answers the court order [S-15], which hardware refresh falls due when [S-21]. That gives the incident-response function something to respond WITH. And your provenance standard needs the same register to know which channels to protect first. So: register wave one and provenance standard ship together as the fast pair; the jurisdiction-of-inference test [S-1] binds at each renewal as the slow, permanent ratchet; fallback capability for services that cannot stop is the expensive tail we cost through the capability fund. Layered, sequenced, no either/or.",
        "refs": []
      },
      {
        "seq": 8,
        "at": "2026-07-10T00:37:59.629Z",
        "participant_ref": "p2",
        "kind": "statement",
        "body": "Synthesis accepted, and Ingrid's challenge changed the shape — the draft will present a first wave of two cheap fast instruments (scoped dependency register plus provenance standard) rather than leading with procurement alone. Full stack for decision: (1) jurisdiction-of-inference test in procurement as the permanent ratchet [S-1]; (2) national register of critical AI dependencies with an incident-response function attached, wave one scoped to top critical services [S-15][S-21]; (3) synthetic-media provenance standard for official communications plus a plain worker usage standard addressing the 48%/56% exposure [S-1]; (4) sovereign fallback requirement for services that cannot tolerate interruption, costed through the capability fund. Framed throughout as resilience against dependency and interference — jurisdictional, not military [G-4].",
        "refs": []
      }
    ],
    "ai_inputs": [],
    "engagements": []
  },
  "korero_segments": [
    {
      "seq": 1,
      "submitted_at": "2026-07-10T00:40:37.395Z",
      "segment_sha256": "2a8a30d1e451c51cb3c32c997b1a8c5ab7e5036ce42e4634ddd9af26c034ba70",
      "entry_range": {
        "from": 1,
        "to": 0
      }
    },
    {
      "seq": 2,
      "submitted_at": "2026-07-10T00:40:37.432Z",
      "segment_sha256": "73572576b3869024916ae2d59ada614cfe3713b4be9b20d896ee9d9587ddd0ee",
      "entry_range": {
        "from": 1,
        "to": 8
      }
    }
  ],
  "receipts": [
    {
      "receipt_version": 2,
      "receipt_id": "gr__usykPZCiP26_SY4",
      "decision_id": "dvr_027DU3mGJkkOAraH",
      "tenant_did": "did:web:devtest.mysovereignty.digital",
      "request_sha256": "f8088bb6ce9ad388627675bfb8a47f416265fc7ae2ac13c53ed542dfc568fbed",
      "action_payload_sha256": null,
      "korero_sha256": "2a8a30d1e451c51cb3c32c997b1a8c5ab7e5036ce42e4634ddd9af26c034ba70",
      "korero_segment_count": 1,
      "reason_codes": [],
      "policy_epoch": "pe_sandbox_v1",
      "decided_at": "2026-07-10T00:40:37.395Z",
      "algorithm": "Ed25519",
      "signer_did": "did:web:devtest.mysovereignty.digital#prf-b98f1a16593cead6271edd93e4b24888",
      "signature": "1a30cfbf0d33843c898fa6862ccc312c76f46a49fbc5dbc8ca92b2de16702440be165eec7d8001996b65c49469a90beca66f218d75c7b59d7a61b3f85372820c",
      "receipt_kind": "record_open",
      "integrity_mode": "tamper_evident",
      "signing_model": "tenant_key"
    },
    {
      "receipt_version": 2,
      "receipt_id": "gr_GH-KO67dLUQZ7vr1",
      "decision_id": "dvr_027DU3mGJkkOAraH",
      "tenant_did": "did:web:devtest.mysovereignty.digital",
      "request_sha256": null,
      "action_payload_sha256": null,
      "korero_sha256": "afa325c203b0d72d200a7c15efe954445f90bbbd8d4d71e1457d46941324d250",
      "korero_segment_count": 2,
      "decision": null,
      "reason_codes": [],
      "policy_epoch": "pe_sandbox_v1",
      "decided_at": "2026-07-10T00:40:37.432Z",
      "algorithm": "Ed25519",
      "signer_did": "did:web:devtest.mysovereignty.digital#prf-b98f1a16593cead6271edd93e4b24888",
      "signature": "833fc9cb5715fa2d6b8f44b36abe13260f9ca57048cd4cc7b0fa0fe95ab7bff5fff65b02737a7656e318752a00dcc2410c700e053685045bc8fe84d95d9f7d06",
      "receipt_kind": "record_segment",
      "integrity_mode": "tamper_evident",
      "signing_model": "tenant_key"
    },
    {
      "receipt_version": 2,
      "receipt_id": "gr_XVsEJtIIg6Xhy0Iv",
      "decision_id": "dvr_027DU3mGJkkOAraH",
      "tenant_did": "did:web:devtest.mysovereignty.digital",
      "request_sha256": null,
      "action_payload_sha256": null,
      "korero_sha256": "afa325c203b0d72d200a7c15efe954445f90bbbd8d4d71e1457d46941324d250",
      "korero_segment_count": 2,
      "decision": null,
      "reason_codes": [],
      "policy_epoch": "pe_sandbox_v1",
      "decided_at": "2026-07-10T00:42:11.404Z",
      "algorithm": "Ed25519",
      "signer_did": "did:web:devtest.mysovereignty.digital#prf-b98f1a16593cead6271edd93e4b24888",
      "signature": "c3d3cebb0bf3b15889f0bd04b36db5100041d7902d835ef3c0bde3cbf39cd97d077f50198d1f0e8d090226b6492983dead727e4465689cd9579b3da738f17308",
      "receipt_kind": "record_seal",
      "integrity_mode": "tamper_evident",
      "signing_model": "tenant_key"
    }
  ],
  "proof_chain": [
    {
      "boundary_crossed": "create",
      "policy_evaluated_by": "sovereignRecordPlugin.create",
      "decision": "allow",
      "caveats_added": [],
      "timestamp": "2026-07-10T00:40:37.400Z",
      "algorithm": "Ed25519",
      "signature": "7b3b57fc5d6150e5015804d6ad837268835d7b98d65e0379c7a385c51ced25c9137bdf9968982158c1b564b2415c0c63cce87a1b78bc09d0f81f2d58a4a81600",
      "signer_did": "did:web:devtest.mysovereignty.digital#prf-b98f1a16593cead6271edd93e4b24888"
    },
    {
      "boundary_crossed": "update",
      "policy_evaluated_by": "sovereignRecordPlugin.update",
      "decision": "allow",
      "caveats_added": [
        "korero",
        "korero.proposal",
        "korero.proposal.description",
        "korero_segments",
        "receipts",
        "korero_sha256",
        "metadata",
        "metadata.origin",
        "metadata.policy",
        "metadata.encryption"
      ],
      "timestamp": "2026-07-10T00:40:37.441Z",
      "algorithm": "Ed25519",
      "signature": "91ff8d50b3ad69e756600139a2412b6a486e3556a7c35dac73569979c0d603ed4652cfc0e4b4d99d79c9e865a01b0341a684f2cb0f4334511ff4f347fae38006",
      "signer_did": "did:web:devtest.mysovereignty.digital#prf-b98f1a16593cead6271edd93e4b24888"
    },
    {
      "boundary_crossed": "update",
      "policy_evaluated_by": "sovereignRecordPlugin.update",
      "decision": "allow",
      "caveats_added": [
        "record_frozen",
        "receipts"
      ],
      "timestamp": "2026-07-10T00:42:11.405Z",
      "algorithm": "Ed25519",
      "signature": "a367cddedb52f5b385e6bdc5591b46d6bd0562548d5ca4ba41766e36048742e60e689fb3d14c0530f5ba7b460e366b29c9e86d086d758cd6b10190dd1f9ff601",
      "signer_did": "did:web:devtest.mysovereignty.digital#prf-b98f1a16593cead6271edd93e4b24888"
    }
  ],
  "origin": {},
  "verification_keys": [
    {
      "id": "prf-b98f1a16593cead6271edd93e4b24888",
      "type": "Ed25519VerificationKey2020",
      "publicKeyMultibase": "z6MkoJz8eJCKmafbRaPYaheAzeRcZxETJxEszRj87YW2FgCd",
      "algorithm": "Ed25519",
      "created_at": "2026-04-22T18:55:57.179Z"
    }
  ],
  "retention": {
    "legal_hold": false,
    "retain_until": null
  },
  "verification": {}
}