The dominant development this week for Australian federal practitioners is the DTA's updated responsible AI policy, now in force, which introduces mandatory APS-wide AI training, internal use-case registers with accountable owners, and pre-deployment impact assessments across fairness, safety, privacy, and related domains — with the first compliance deadline arriving in June 2026. Alongside this, the Good Ancestors January newsletter draws together several threads worth tracking: the Australian AI Safety Institute is standing up its founding team under resource constraints, the Productivity Commission has confirmed its preference against AI-specific regulation as a first resort, and the ACCC has flagged agentic AI as an emerging collusion risk. For practitioners thinking beyond immediate compliance, technical reporting this week notes that decentralised AI training capacity is scaling rapidly, a development with longer-term relevance to questions of AI access, provenance, and supply chain assurance. Returning from the break, the immediate priority is assessing agency readiness against the DTA policy's phased implementation timeline.
The DTA has released an updated Policy for the responsible use of AI in government, effective 15 December 2025, applying to all non-corporate Commonwealth entities. Key additions include mandatory foundational AI training for all APS staff, a requirement to maintain an internal register of in-scope AI use cases with assigned accountable owners, and pre-deployment AI impact assessments covering fairness, safety, privacy, transparency, security, and human-centred values. A new AI impact assessment tool supports compliance. The first mandatory requirement commences 15 June 2026, with remaining requirements taking effect December 2026, providing a phased implementation runway for agencies.
Implications
ImplementAgencies could establish or update internal AI use case registers, assign accountable owners, and integrate pre-deployment AI impact assessments into their governance processes ahead of the June 2026 deadline.
ImplementHR, learning and development, and AI governance teams could plan and roll out mandatory foundational AI training for all staff in line with the December 2026 compliance timeline.
ConsiderAgencies may want to assess current AI incident reporting pathways and staff-facing AI safety concern mechanisms against the new minimum practice requirements to identify gaps.
Implications are AI-generated. Starting points, not advice.
AU(undated)Good Ancestors – AI Policy & Governance Newsletter
This January 2026 newsletter from Good Ancestors consolidates several significant Australian AI policy developments. Australia's AISI is recruiting its founding team amid expert warnings about underfunding ($7.5M per year vs a recommended $50M+) and bureaucratic culture risk. The Productivity Commission has finalised its recommendation against AI-specific regulation as a first resort, while the ACCC has flagged emerging risks from agentic AI including collusion. MYEFO funding of $166 million for GovAI Chat and a Grok deepfake investigation by the eSafety Commissioner add urgency to both AI delivery and harm governance agendas across the APS.
Implications
MonitorAgencies involved in AI governance could track AISI founding team appointments and early operational priorities, as these will shape the institute's culture and regulatory posture.
ConsiderPolicy and procurement teams may want to assess how GovAI Chat's $166 million investment interacts with existing agency AI tooling decisions and whole-of-government AI strategy.
ConsiderAgencies with consumer-facing or automated decision-making functions could review the ACCC's agentic AI snapshot against their current risk and assurance frameworks.
Implications are AI-generated. Starting points, not advice.
This edition of Import AI covers two research developments. First, Meta's KernelEvolve system demonstrates AI-driven automation of hardware kernel design, achieving significant infrastructure efficiencies at hyperscale and illustrating how AI is increasingly used to accelerate its own development pipeline. Second, an Epoch AI analysis of decentralised AI training finds compute scale growing at 20x per year—far outpacing frontier centralised training's 5x growth—though decentralised runs remain roughly 1000x smaller than frontier models. The newsletter frames decentralised training as a political technology that could broaden access to frontier-scale AI beyond the current small cluster of US and Chinese tech companies, with potential implications for academic, government, and non-commercial AI development globally.
Implications
MonitorDISR, AISI, and strategy teams may want to monitor decentralised AI training trends, as rapid compute growth could affect assumptions about who develops frontier-capable models and how Australian policy responds.
MonitorAgencies tracking AI supply chain and infrastructure risk may want to watch whether AI-automated kernel development accelerates capability timelines in ways that affect existing risk assessments.
Implications are AI-generated. Starting points, not advice.