EN18286
🟢 Why METR matters? METR exposes the same system‑level physics EN18286 formalises: alignment failures, monitoring gaps, and acceleration outrunning oversight. Their proposal is not alignment theory; it is evidence generation for loss‑of‑control risk inside companies. 🟢 The public evidence gap * “The public currently lacks the visibility into the internal
Health
🟢 Why clinical‑grade AI fails long before the model does? The recent surge in AI‑health discussions shows a familiar pattern: Healthcare leaders want acceleration, BUT the systems beneath clinical AI are not built for it. The friction isn’t in the model. It’s in the DATA, DECISIONS, and
AiKids
🟢 The human layer: the 95% Humans don’t form their worldview from structured reasoning. They form it from identity anchors — early phrases that lodge deep and stay for decades. These anchors become: - worldview filters - behavioural defaults - moral baselines - decision heuristics - lifelong beliefs For some, the anchors are governance ideals: - “By
AIGP
🟢 Overview EN18286 - The first harmonised STD under the EU AI Act. * Harmonisation Is the Easy Part. The real test is whether providers can operationalise it. Harmonisation under Implementing Decision (EU) 2026/2274 finally gives providers a clear presumption of conformity pathway for Article 17. For the first time, high‑risk
ACW
🟢 Overview * GenAI didn’t just accelerate creation — it collapsed the middle of the process. * Open‑web telemetry shows three structural signals defining a new organisational cost centre inside engineering teams: The Cleanup Economy. 🟢🟢 The Cleanup Economy 🟢1️⃣ The 8× Code Duplication Shock * Telemetry across 623M code changes shows an eightfold
AiKids
Introduction The West is finally waking up to the governance physics of synthetic relationships. * On 17 September 2026, the European Commission unveiled the draft EU KIDS Act * The real structural story sits in Article 14: banning emotional‑dependency AI for minors * CAC reached the same diagnostic conclusion through telemetry: retention‑
AiKids
Western Enforcement Paradox (AiKids Section Update) The EU KIDS Act is the strongest Western attempt to redesign digital spaces for minors. It expands “Social Media” into “Social Media+”, explicitly covering online games, AI chatbots, and AI companions. * The Act flips the burden-of-proof: platforms must prove safety before deployment.
ACW
What is it? Tech frameworks define identity proofing, runtime safety and accountability roles but do not fully govern the human owner of automated systems. * Automation in tech environments creates decisions that require clear ownership boundaries * The ACW Model (Accountability CounterWeight) provides the missing structure for identity, runtime and strategic accountability
ACW
What is it? Health systems increasingly rely on automation for triage, alerts and clinical decision support * Existing frameworks govern elements of ownership but do not provide a full model for the human owner of automated clinical systems * The ACW Model (Accountability CounterWeight) provides the missing structure for identity, runtime and
ACW
What is it? Finance is one of the most automated sectors, with decisions shaped by systems across risk, reporting and capital processes * Existing frameworks govern parts of ownership but do not provide a complete model for the human owner of automated systems * The ACW Model (Accountability CounterWeight) provides the missing
RACinform
Overview AI Governance by #Sectors Series. People are searching for plain‑English clarity about how AI is impacting their work and their sector. The demand is clear, and the gaps are real. The #Sectors series maps how the UK’s ~33 million workers are distributed across major sectors — originally published
RACinform
Overview The Cyberspace Administration of China (CAC) is one of the most influential digital governance bodies in the world. Its regulatory posture shapes: * platform behaviour * algorithmic transparency * data flows * AI deployment boundaries * content governance This newsletter examines CAC as a case surface for understanding how state‑level guardrails influence AI