AI in the Not-for-Profit Sector: Opportunity, Oversight and the Human Touch

At the recent update for members of the Australian Institute of Company Directors, three main topics were the subject of presentations: climate governance and mandatory reporting, data governance and cyber security, and artificial intelligence.

The Australian Charities and Not-for-profits Commission (ACNC) issued guidance to its sector constituents about the use of AI, but in the interests of transparency, it also issued a statement about its own use of AI in August 2025.

In the guidance issued to charities, the ACNC noted examples of where charities are already using AI, including in learning outcomes, fundraising and addressing the impacts of climate change.

Charities are also making broad use of suites of AI-based tools that assist with labour-intensive, back-office functions such as board paper preparation and diary management, or even image or graphics generation.

In a sector facing the challenge of chronic under-funding after years of funding cuts, the potential for AI to fill some of the resourcing gaps is tantalising to many.

But we also know that this particular technological “advancement” comes with a host of challenges, not least of which are data management and privacy concerns.

Deloitte Australia was forced to refund $440,000 to the Australian government after a $500,000 report was found to be riddled with apparently AI-generated errors, including fabricated quotes and research.

Where does this leave the jobs market and what can the Pro Bono Salary Survey 2025 tell us about AI?

Survey respondents were asked whether they use AI in their current role and results show us that uptake is rising and reiterating that this is an issue clearly at the forefront of everyone’s minds.

Equally, for anyone worried they’re going to lose their job to AI, while that may well be a long-term possibility, this fairly nascent technology is unlikely to push anyone out of a job this week or next while the horror stories abound.

Don’t get me wrong, there’s no question AI has the potential to fill some resourcing gaps. However anyone familiar with using the technology will tell you that right now, all the functions being served by AI still need human oversight and checking, lest the robots do you dirty.

When I was testing ChatGPT to see what functions it could help me with, I provided it with a transcript and asked it to pull out relevant quotes on specific topics. When I went back to cross-reference those quotes, I realised it had provided me with a long and comprehensive list of made-up gobbledygook being presented as fact (something I probably should have mentioned to Deloitte Australia).

An HR manager I know who works at a very large organisation, put a position description into ChatGPT, for the purpose of checking the text for any spelling or grammar errors. The job was later filled internally and not advertised and the PD left unpublished, but content from the document mysteriously started appearing online soon after.

In aiming to give insights into its own use of AI, the ACNC’s August statement offers a best-practice model that charities can also use when assessing their own functions, including appropriate trust, governance, risk assessment and monitoring.

What tasks are AI-appropriate? What are not? How can we ensure the information we’re entering is safe? How can we ensure the information we are receiving is safe?

In America, more than half of not-for-profit organisations are using AI but less than 10 per cent have an AI policy.

It will take a long time before AI is at the point where it can do without substantial human oversight and intervention and it’s likely it will need issue-specific governance frameworks and management for the foreseeable future.

Ultimately, the crux of this is that employers need staff who understand the technology, its possibilities and limitations and that means that employees with those skills will be highly sought-after.

These skills will be important across many areas and functions, but keeping in mind the data quoted above from America about the dearth of organisational AI policies in place, these skills will be immediately crucial in areas like risk management and mitigation, as well as technology, data privacy and legal.

It also means that if you’re thinking about a career change or upskilling, voilà — this is a space worth exploring. Or at the very least, it’s a valuable set of skills to consider adding to your toolkit.