Agile InsightsBriefing
Reference 14 cover: a closed reference volume with headphones, titled PSPF + ISM.

Reference 14 · Primer · About 6 minutes

Security for government AI: the PSPF, the ISM and the frontier AI era

What the protective security framework and the Information Security Manual now ask of agencies and AI suppliers.

#rules#security

Download the MP3 · Linda, an AI voice, reads a script written from the sources below.

In this episode

01Two rulebooks for AI securityThe Protective Security Policy Framework from Home Affairs sets the obligations for non corporate Commonwealth entities. The Information Security Manual from the Australian Signals Directorate holds the technical controls.
02Basics before frontier AIA May 2026 advisory said agencies do not need the most advanced AI to stay protected. Get patching and hardening right first. Since 2025, AI must go through technology system authorisation.
03A legacy stocktake by March 2027A September 2026 direction, as reported, requires a legacy technology stocktake and risk plan by the end of March 2027, with reduction targets and faster patching.
04New controls for AI agentsIndustry summaries of the 2026 manual updates report controls for agent identity, an agent register, task-specific grants, central logs of tool calls and human approval of high impact actions.
05What to do nowStart the legacy stocktake, give every agent its own identity and a register entry with an owner, and, if you supply AI, prepare evidence for the agent controls before an assessor asks.
Sources and links (8)
PSPF Release 2026 now available · protectivesecurity.gov.au
Home Affairs on what the 2026 release adds, including AI use training.
PSPF publications library · protectivesecurity.gov.au
Where the advisories and directions are published.
What's new in the PSPF · KWM, 10 Sep 2025
King & Wood Mallesons on the 2025 release, including AI authorisation.
Fix security basics before frontier AI · iTnews, 28 May 2026
Reporting on Policy Advisory 001-2026.
Gov-wide legacy system stocktake · iTnews, 30 Sep 2026
Reporting on Direction 002-2026 and its March 2027 deadline.
What changed in the June 2026 ISM · idm.net.au, 27 Aug 2026
A practitioner summary of the June AI controls.
ASD ISM September 2026 update · Cipher Projects
A practitioner summary of the agent identity, register and tool-call controls.
2023 to 2030 Australian Cyber Security Strategy · homeaffairs.gov.au
Home Affairs' strategy page, including Horizon 2.

Written from Home Affairs and ASD material and practitioner summaries, checked on 11 November 2026. The ISM control wording comes from secondary summaries. Check the current text before relying on it.

Our fact check · 13 claims: 8 confirmed, 4 reported

Before we record, every claim the hosts make is checked against the sources (this table was last updated 11 November 2026). Confirmed means a primary source or two independent reports; Reported means one party says so; Partly right, Disputed, Unverified and Not yet known mean what they say, and the hosts say so on air. This is the table as we left it.

  1. ConfirmedThe Protective Security Policy Framework belongs to Home Affairs, is mandatory for non-corporate Commonwealth entities, and reaches suppliers through contractsKing & Wood Mallesons, 10 September 2025
  2. ConfirmedThe Information Security Manual comes from the Australian Signals Directorate and is updated quarterlyAustralian Signals Directorate, cyber.gov.au
  3. ConfirmedSince the 2025 release, agencies must put AI technologies through technology system authorisation, wherever they are hostedKing & Wood Mallesons, 10 September 2025
  4. ConfirmedA 2025 direction barred DeepSeek productsKing & Wood Mallesons, 10 September 2025
  5. ConfirmedThe 2026 release adds training on AI useHome Affairs, PSPF Release 2026 now available
  6. ReportedA May 2026 advisory said agencies do not need the most advanced AI models to stay protected, and should fix the basics firstITnews, 28 May 2026; govtechreview, 1 June 2026
  7. ReportedA September 2026 direction requires a legacy technology stocktake, a risk plan, reduction targets and faster patchingITnews, 30 September 2026
  8. ReportedThe stocktake and plan are due by the end of March 2027ITnews, 30 September 2026
  9. Security and AI teams will compete for the same peopleOur analysis; —
  10. ReportedThe 2026 manual updates add controls for agent identity, an agent register, task-specific grants, untrusted retrieved content, central tool-call logs and human approval of high impact actionsCipher Projects, September 2026 update; idm.net.au, 27 August 2026
  11. ConfirmedAssessors under the Infosec Registered Assessors Program test systems against the manualAustralian Signals Directorate
  12. ConfirmedThe DTA's AI Reference Architecture defers to existing guidance on security without naming itAIRA v1.0, p. 15 (that this means the framework and the manual for government is our reading)
  13. ConfirmedHorizon 2 of the Cyber Security Strategy launched in June 2026 and includes crisis planning for major AI incidentsHome Affairs, 2023 to 2030 Australian Cyber Security Strategy

How this maps to what we do

The rules that apply to AI in Australia and for Australian organisations with European exposure, and the practical work each one calls for.

Standard or rule
What an organisation needs
Where we help
Privacy Act: automated decision-making transparency (from 10 December 2026)Episode: Is your AI making decisions about people? What the Privacy Act now asks of you
A list of the automated decisions made about people, including those buried in vendor software, and the policy wording that describes them.
AI Risk & ComplianceUse Case Portfolio & AssessmentAI Governance Framework Alignment
Australia's AI guidance: the six essential practices (with VAISS, SafeAI-Aus and NAIC)Episode: Australia's AI guidance: the six essential practices
A baseline against the six practices, then a prioritised roadmap to close the gaps.
AI DiagnosticAI Governance Framework AlignmentAI Strategy Build
Each AI system classified by risk tier, with the obligations and dates that apply to it, including for non-EU organisations.
AI Risk & ComplianceUse Case Portfolio & AssessmentAI Governance Framework Alignment
Australian sector overlays: government, health, finance, education, disability, aged care, not-for-profitEpisode: Australian sector rules for AI: what applies to your industry
The sector's own obligations mapped onto one set of AI controls, so that nothing is done twice.
AI Risk & ComplianceAI Governance Framework Alignment
A review of worker monitoring and automated decisions about people for safety and psychosocial risk.
AI Risk & ComplianceAdoption & Engagement

These are Agile Insights AI Strategy & Governance offerings. Which of them you need depends on where your organisation starts.

Want to know more?

Sign up for our podcast at agile-insights.com.au.

Read the transcript

Linda The security rules for government AI changed several times in 2026. The newest change gives agencies a deadline at the end of March 2027. By the end of this episode, you will know the two rulebooks involved and what each now asks. You will also know what a security assessor will ask an AI supplier.

Linda First, what they are. The Protective Security Policy Framework, the P S P F, belongs to the Department of Home Affairs. It is mandatory for non corporate Commonwealth entities, and it reaches suppliers through their contracts. Underneath it sits the Information Security Manual, the I S M, from the Australian Signals Directorate. The framework sets the obligations. The manual is the catalogue of technical controls, and it is updated every quarter.

Linda The problem they solve is that AI widens the attack surface while also speeding up attackers. An agency with old, unpatched systems is exposed either way. Picture an AI agent that can read email and call other tools. Suppose it has no identity of its own and keeps no log. Then nobody can say who acted, or stop it quickly.

Linda Who uses them? Chief security officers and chief information security officers own the obligations. Security assessors under the Infosec Registered Assessors Program, called I RAP, test systems against the manual. Procurement teams write the framework into contracts. And suppliers of AI products meet both when they go through an assessment.

Linda So what is in them? Start with the framework. Since its 2025 release, agencies must put AI technologies through technology system authorisation, wherever they are hosted. A 2025 direction barred DeepSeek products. The 2026 release adds training on AI use. In May 2026, Home Affairs issued an advisory on the frontier AI era. Its message was blunt. You do not need the most advanced AI to stay protected. Get the basics right first, starting with patching and hardening.

Linda Then, in September 2026, came a binding direction on AI enabled risks. As reported, agencies must take stock of their legacy technology and write a plan to manage its risk. They must also set reduction targets and patch critical flaws faster. The stocktake and plan are due by the end of March 2027. That lands beside the AI policy's own dates, so security and AI teams will compete for the same people.

Linda Now the manual. Industry summaries of the June and September 2026 updates report a new set of AI controls. Each AI agent gets its own identity, separate from people and from other agents. Agents sit in a register with an owner, a purpose and the tools they can reach. A tool call needs both the user's access and a task specific grant for the agent. Content an agent retrieves stays untrusted. Every tool call is logged centrally. And a human approves high impact actions before they run.

Linda How does this connect to the rest of the stack? The D T A's AI Reference Architecture says it will defer to existing guidance on security, without naming it. For government, that guidance is the framework and the manual. The national Cyber Security Strategy sits above both. Its second horizon, launched in June 2026, adds national AI work such as crisis planning for major AI incidents.

Linda Getting started. First, start the legacy stocktake now, because March 2027 is close. Second, give every AI agent its own identity and an entry in a register with a named owner. Third, if you supply AI products, prepare evidence for those manual controls before an assessor asks. The primary sources are on protective security dot gov dot A U and cyber dot gov dot A U. Both are linked in the show notes.

Linda Now, currency. As at November 2026, the framework is at its 2026 release, and the manual's latest update is from September. The manual changes every quarter, and the control wording here comes from industry summaries. Check the current text before you rely on any detail here.

Linda In short, government AI security starts with the basics. Then every agent gets an identity, a log and a human brake. If you'd like to know more, sign up for our podcast at agile-insights.com.au.

More on #rules

Reference 8 · PrimerISO/IEC 27001 and 27701: security and privacy, for AI tooReference 15 · PrimerThe quiet obligations: records, data ethics, digital experience and cloudReference 13 · PrimerThe Commonwealth AI rulebook: the Policy, the technical standard and the APS AI PlanReference 12 · PrimerThe DTA AI Reference Architecture: one blueprint for government AIEpisode 21 · DiscussionTrust cannot be rushed: how much authority will you give AI?Reference 11 · PrimerAustralian sector rules for AI: what applies to your industry

Every episode tagged #rules

All of the reference series (16) · Follow AI Governance Reference in Apple Podcasts or any podcast app

Everything from Agile Insights: briefings, Making You Think and the reference series