docs: Update Potaua email draft — add contact details and introducer name

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This commit is contained in:
TheFlow 2026-02-18 09:17:38 +13:00
parent 772f32008b
commit 5d6bb6482b

View file

@ -114,7 +114,7 @@ Balcairn, North Canterbury
## 4. Potaua Biasiny-Tule — Digital Natives Academy / UNESCO HILEG-ELT
**To:** [via introduction — confirm address]
**To:** potaua@internetnz.net.nz
**Subject:** Request for critical review — Taonga-Centred Steering Governance (STO-RES-0010)
---
@ -123,7 +123,7 @@ Tēnā koe Potaua,
My name is John Stroh. I am a retired technologist based in North Canterbury, working on a small research programme concerned with sovereign AI deployment for communities in Aotearoa, documented at agenticgovernance.digital.
I am writing to you — via [introducer's name]'s kind introduction — because we have produced a draft paper that attempts to address architecturally what you have been advocating for publicly: that tikanga should shape how AI operates for and with Māori. The paper, "Taonga-Centred Steering Governance: Polycentric Authority for Sovereign Small Language Models," proposes a governance architecture in which steering vectors — the mathematical instruments used to adjust language model behaviour at inference time — are governed polycentrically, with iwi and community authorities operating as co-equal peers to the platform operator rather than as downstream consumers of its corrections.
I am writing to you — on the basis of TeRata Kikairo's recommendation— because we have produced a draft paper that attempts to address architecturally what you have been advocating for publicly: that tikanga should shape how AI operates for and with Māori. The paper, "Taonga-Centred Steering Governance: Polycentric Authority for Sovereign Small Language Models," proposes a governance architecture in which steering vectors — the mathematical instruments used to adjust language model behaviour at inference time — are governed polycentrically, with iwi and community authorities operating as co-equal peers to the platform operator rather than as downstream consumers of its corrections.
The central argument is that some domains of cultural knowledge are structurally outside the platform operator's authority to define or correct. Your work on UNESCO's High-Level Expert Lead Group on the Governance of Ecosystem-Level Transformation in AI addresses this at the ecosystem level — the question of how AI governance structures should accommodate, rather than subordinate, indigenous authority. Our paper arrives at a similar question from a different direction: what does a technical architecture look like that actually implements polycentric authority at the inference layer?
@ -146,6 +146,13 @@ I recognise that your commitments are substantial and that unsolicited requests
Ngā mihi nui,
John Stroh
agenticgovernance.digital
Balcairn, North Canterbury
My Digital Sovereignty Ltd
M +64272418020
Research Website: https://agenticgovernance.digital
Commercial Website: https://mysovereignty.digital