Artificial Intelligence (AI)

Responsible AI Governance for Australian Not‑for‑Profits

1 March 2026

Robotic hand and human hand reaching out, symbolizing AI and human collaboration for responsible AI governance.

Responsible AI governance for not-for-profits is increasingly discussed across Australia’s not‑for‑profits, particularly in relation to AI governance policy and board oversight of AI.

Board papers, conference agendas and funding discussions often assume that AI is already widely embedded in community services. However, recent Australian research paints a more nuanced picture.

While many NFPs are experimenting with AI at an organisational level, frontline use in social services remains limited, cautious and highly contested. For boards, responsible AI governance starts with understanding this difference – and governing accordingly.

Is AI Actually Being Used in Social Services?

Australian evidence shows that AI use in not‑for‑profits is uneven.
Sector‑wide studies, including the 2025 Infoxchange Digital Technology in the Not‑for‑Profit Sector Report, indicate that many organisations are now using generative AI – particularly for administrative, communications, fundraising and reporting tasks.

At the same time, research focused on social work and child and family services shows that frontline AI use remains limited, highly constrained and ethically contested, reflecting the higher risks involved in direct service delivery.

See:

  • Research by the Institute of Community Directors Australia (ICDA) on AI use by NFP boards and organisations
  •  A 2025 evidence review and workforce survey by the Centre for Excellence in Child and Family Welfare on AI in social work and child and family services
  • 2025 Digital Technology in the Not-for-Profit Sector report based on insights of 800+ NFPs across data, cyber, digital transformation and AI.

At first glance, these sources appear to say different things. In practice, they are describing different parts of the sector.

What the ICDA Research Tells Us

ICDA’s research focuses largely on boards, executives and organisational staff across the NFP sector.

Key findings include:

  • A growing number of NFPs report using or trialling generative AI tools, particularly for administrative, communications, governance and policy tasks
  • AI use is often informal and uneven, driven by individuals rather than organisational strategy
  • Formal AI policies and governance arrangements are rare, even where AI tools are being used
  • Boards are aware of AI risks but governance is generally lagging behind experimentation.

This research shows that AI is entering NFPs primarily through lower‑risk corporate and governance functions, not through direct service delivery.

What the Social Work Research Tells Us

The Centre for Excellence in Child and Family Welfare research focuses on frontline practitioners in child, family and social services – including family violence, mental health and safeguarding contexts.

Its findings are notably different:

  • 66% of surveyed practitioners had never used AI at all
  • Only 8% reported using AI in their work
  • Many organisations explicitly prohibit AI use, citing privacy and confidentiality concerns

Practitioners expressed strong concerns about:

  • loss of human nuance and relational practice
  • bias, misinformation and lack of explainability
  • deskilling and erosion of professional judgement
  • misuse of AI by organisations to increase workloads.

At the same time, there was cautious optimism about very limited uses of AI to reduce administrative burden like drafting notes or speech‑to‑text, provided strong governance and human oversight are in place.

Why These Findings Are Not Contradictory

These two evidence bases are not saying opposite things. They are describing a two‑speed reality:

  • At the organisational level, NFPs are experimenting with AI for governance, communications and efficiency
  • At the practice level, particularly in human‑centred and high‑risk services, AI use is deliberately constrained.

This difference exists for good reasons:

  • Social services involve high stakes decisions, duty of care and trauma‑informed practice
  • Professional ethics emphasise human judgement, empathy and context
  • The risks of error, bias or privacy breaches are far greater in frontline settings
  • Practitioners are rightly cautious about tools that may undermine trust or safety

For boards, recognising this distinction is essential to credible governance.

What this means for governance

Taken together, Australian evidence points to a consistent risk pattern. AI is becoming normalised in not‑for‑profit organisations for administrative, communications and reporting tasks, while remaining deliberately constrained in frontline, human‑centred services.

At the same time, governance arrangements have not kept  pace. The 2025 Infoxchange Digital Technology in the Not‑for‑Profit Sector Report found that while 67% of not‑for‑profits are using generative AI in some form, only 14% have an AI policy or guideline in place, and half cite privacy, security, data sovereignty or ethical risk as their primary concern.

This gap between AI use and formal governance is where board responsibility now sits.

What Responsible AI Governance Means for Boards

Responsible AI governance is not about pushing adoption. It is about setting clear boundaries, safeguards and accountability that reflect real‑world use and risk.

In practice, this means boards should ensure:

  • Visibility –  there is sufficient oversight and understanding of where AI is (and is not) being used across the organisation
  • Policy clarity – there is explicit guidance in place on acceptable and prohibited AI use, including public generative AI tools
  • Human‑led decision‑making –  policy clearly states that critical decisions must remain person‑led
  • Privacy and transparency – that AI inputs and outputs must be aligned with privacy requirements (ie the Privacy Act 1988)
  • Workforce confidence –  the organisation facilitates training and guidance for staff that support informed, ethical practice instead of silent experimentation (ie the likely alternative)
    Proportional governance – stronger controls are in place for higher‑risk contexts such as family violence, child protection and mental health

Queensland’s increasing emphasis on structured AI governance in the public sector further reinforces these expectations for funded NFPs.

Evidence‑Led Governance, Not Assumption‑Led Adoption

Taken together, Australian research sends a consistent message:

AI should be approached as a supporting tool, not a decision‑maker – and only where it aligns with professional values, ethical practice and community trust.

For boards, the governance task is not to assume AI use, but to govern thoughtfully where it exists, and deliberately where it does not.

Clear, current policies are one of the most practical ways boards can demonstrate this oversight.

The Policy Place’s online policy service supports Australian not‑for‑profits with governance‑ready policy frameworks that reflect emerging AI risks, workforce realities and regulatory expectations.

References and Further Reading

false