Resources

Practical resources for trusted AI adoption.

AAIIG resources will focus on implementation reality: what buyers should ask, what providers should evidence and what responsible adoption looks like in Australian conditions.

Guide

Code of Practice

AAIIG's code of practice gives members a plain-English baseline for developing, buying and deploying AI in ways that are accountable, safe, transparent and useful.

Draft v1 baseline

The guide translates Australia's Voluntary AI Safety Standard and the National AI Centre's six essential AI practices into practical member commitments. It is designed for SMEs, advisers, product teams, procurement leads and executives who need a clear starting point before heavier assurance or accreditation work begins.

Member commitments

  • Assign accountable owners for every AI use case and maintain an AI system register.
  • Screen each use case for impact, affected stakeholders, legal obligations and unacceptable risk.
  • Document data sources, data rights, privacy settings, security controls and model or supplier provenance.
  • Disclose meaningful AI use to users and maintain clear channels for feedback, challenge and redress.
  • Test systems before release, monitor performance after release and keep evidence of decisions and changes.
  • Require suppliers to provide enough transparency for procurement, risk review and ongoing assurance.
  • Consider fairness, accessibility, inclusion and Indigenous Data Sovereignty where people or communities may be affected.
  • Share incident learnings through AAIIG working groups where doing so can improve sector practice.

Evidence base

1

The Australian Government's Voluntary AI Safety Standard sets 10 guardrails covering accountability, risk management, data governance, testing, human oversight, transparency, contestability, supply chain transparency, record keeping and stakeholder engagement.

2

The National AI Centre's 2025 Guidance for AI Adoption simplified this into six essential practices for responsible governance, risk management, testing, supply chain controls and human oversight.

3

The 2024 Australian Responsible AI Index found a mean responsible AI maturity score of 44 out of 100, with only 8% of surveyed organisations in the leading stage.

Source material

Playbook

AI Development and Insurance Playbook

A practical playbook for AI developers, adopters and advisers who need to show that an AI system is controlled, monitored and fit for insurance, procurement and board review.

Implementation draft

The playbook connects product delivery with risk evidence. It helps teams prepare the artefacts insurers, customers, boards and regulators increasingly expect: ownership, data provenance, model behaviour evidence, security controls, supplier transparency, incident planning, and fallback arrangements for critical operations.

Project gates

  • Discover: define the business outcome, affected stakeholders, decision rights, risk appetite and insurance context.
  • Design: document data rights, privacy obligations, model choices, supplier dependencies, IP boundaries and expected limitations.
  • Build: use secure engineering controls, versioning, evaluation datasets, human review points and release authority.
  • Assure: run bias, safety, security, performance and misuse testing proportionate to the impact of the use case.
  • Insure: review professional indemnity, cyber, technology errors and omissions, directors and officers, and contractual liability settings for AI-specific gaps.
  • Deploy: publish user disclosures, escalation paths, monitoring thresholds, rollback triggers and business continuity fallbacks.
  • Operate: maintain logs, change records, incident reports, supplier notices and ongoing monitoring for drift, bias and control failure.
  • Retire: decommission models, revoke access, preserve records and confirm data retention or deletion obligations.

Evidence base

1

APRA's April 2026 AI letter observed that adoption across banks, insurers and superannuation trustees is moving beyond experimentation, while governance, assurance and operational resilience are not always keeping pace.

2

APRA expects AI lifecycle ownership, inventories of tools and use cases, human involvement for high-risk decisions, staff training, supplier transparency and continuous monitoring proportionate to criticality.

3

Insurance risk is no longer only a cyber question. AI can create professional negligence, privacy, IP, discrimination, operational resilience, supplier concentration and board oversight exposures.

Source material

Report

State of the Australian AI Industry

A data-led industry briefing that frames Australia's AI opportunity as both a maker and taker of AI: strong research and adoption, with a clear need to lift commercialisation, skills and trusted deployment.

Launch-year briefing

The report turns public ecosystem research into an AAIIG publishing agenda. It identifies where the industry is growing, where evidence is still thin, and what AAIIG can track annually through members, chapters, working groups and the public directory.

Current signals

  • Australia's AI ecosystem sample includes 1,533 AI companies, including 1,121 private companies and 412 public companies.
  • Private AI firms are predominantly small: 85% of the sampled private companies employ fewer than 50 staff.
  • The NAIC and CSIRO identified 25 geographic AI clusters containing 858 geocoded firms, with Melbourne CBD the largest cluster at 188 companies.
  • AI-related Australian patents rose from 170 in 2015 to 629 in 2024, while AI-related research publications more than doubled.
  • Australia produced 93,302 AI-related research publications and 4,075 AI patents between 2015 and 2024, showing a persistent research-to-commercialisation gap.
  • The National AI Plan reports more than 1,500 AI companies, $700 million in private AI firm investment in 2024 and AI-skilled worker demand tripling since 2015.
  • Responsible AI maturity remains uneven: the 2024 Responsible AI Index found most surveyed organisations in the developing or implementing stages.
  • AAIIG should track adoption maturity, procurement readiness, investment pathways, workforce capability, regional clusters and member evidence each year.

Evidence base

1

The 2025 NAIC and CSIRO ecosystem update describes Australia as a dual-track AI taker and AI maker, applying global foundation models while developing targeted domestic capability.

2

The National AI Plan frames AI as a national productivity, capability and safety priority, with goals to capture the opportunity, spread benefits and keep Australians safe.

3

AAIIG's role is to add practical industry signal: member case studies, implementation barriers, assurance gaps, chapter-level capability and supplier evidence that public datasets cannot fully capture.

Source material

Member value

The member library will grow from working group output.

The member library will support resources, event recordings, working group notes, policy consultation material and future accreditation content as membership opens.

Resource themes

Codes of practice Procurement guidance AI assurance Implementation playbooks Incident learning

Launch targets

The resource program should become measurable.

AAIIG's early success measures include a national footprint, practical standards material, accreditation pathways and an annual evidence base for the industry.

1

National footprint with state and territory chapters

These targets give AAIIG a clear publishing roadmap as public launch content grows into member operations.

2

Code of Practice and practical deployment playbooks

These targets give AAIIG a clear publishing roadmap as public launch content grows into member operations.

3

Accreditation and CPD pathways for practitioners and suppliers

These targets give AAIIG a clear publishing roadmap as public launch content grows into member operations.

4

Annual State of Australian AI Industry report

These targets give AAIIG a clear publishing roadmap as public launch content grows into member operations.