Our approach
Governance does not produce capable staff. Capable staff produce governance.
Not For Humans builds AI-capable humans in not-for-profits, NGOs, and community organisations, through behavioural awareness and practice.
Most organisations try to build AI governance from the top down. Write a policy, run a session, expect behaviour to follow. It rarely does, and not for want of good intentions. Behaviour does not change because a policy asked it to. It changes because people have built the capability to do things differently.
So we build in the opposite order. Individual capability first. Policy last.
AI tools change weekly, not yearly. A policy written around today’s tools describes a target that will have moved by the time it is approved. A team with real capability does not have that problem. The same judgement applies to a tool they have never seen before, because the capability sits with the people rather than the document.
Photo: Effydesk / Unsplash
The build order
The same model, read in the other direction.
On the Science page this model reads from the outside in, to show where risk sits. Here it reads from the inside out, because that is the order a response has to be built in. Start at the centre with human judgement. Work outward. Policy arrives last, on top of capability that already exists rather than standing in for capability that does not.
Not For Humans, 2026. Read inward out: build Layer 1 first, Layer 4 last. Capability first, policy last.
Photo: Krismas / Unsplash
Build individual capability first.
Disclosure becomes expected rather than merely permitted, and a named person stands behind every AI-assisted output.
AI Authorship Statement →Build team culture, not just individual habits.
A shared vocabulary lets a colleague name a pattern without it landing as criticism of the person.
Human Behaviour Cards →Build the capability to name bias without judgement.
Structured practice with feedback, rather than a lecture explaining that bias exists.
Leadership and AI Reflection Workbook, Not For Humans Behaviour Lab →Build review into a process, not a good intention.
A structure that exists whether or not anyone remembers to be careful that day.
AI in Use Audit, AI Accountability Statement →Build policy on top of capability, not instead of it.
A small number of durable principles, and a team capable of applying them to whatever turns up next.
AI Governance Model, Policy Template, Action Plan →“I have watched staff make colleagues feel small for using AI. It is quick, it is public, and it is rarely unkind by intention. What I notice every time is the same thing: everyone else in the room deciding, silently, never to mention their own use.
That is the moment AI goes underground. Not through a policy failure. Through one raised eyebrow that everyone else read correctly.”
Adrian Hubble, Founder, Not For Humans
Reduction in biased decision-making after a single structured training session, with effects still measurable two months later
Reduction in surgical deaths from a two-minute structured checklist. Not new equipment or new skill. Structure replacing memory under pressure.
Photo: Vitaly Gariev / Unsplash
Each stage depends on the one before it. A team cannot build honest review on top of a culture where naming bias feels like an accusation. A policy cannot govern behaviour nobody has had the chance to practise noticing in themselves.
Where should we start?
Not every organisation needs to address all four layers at once. Find the row that sounds like you.
Where to start depends on where the biggest gap is. Tap the row that describes your organisation.
The honest starting point. Before any of this, the most useful thing a leadership team can do is have each leader assess, alone, where they think the organisation sits across the four layers. Then compare. The disagreements between leaders about where the organisation sits are usually more informative than the agreements.
How We Work
Our approach rests on four pillars.
Not For Humans does not provide policies, compliance checklists, or top-down frameworks. We provide practical infrastructure: the shared language, pause points, and accountability structures that make good judgment possible in the environments community organisations actually operate in.
Sector Evidence
AI tools are being adopted at twice the rate of last year across the not-for-profit sector. Yet only 14% of not-for-profits have an internal AI policy or governance framework in place. The gap between adoption and governance is widening.
Infoxchange Digital Technology in the Not-for-Profit Sector Report 2025 ↗Capability and Confidence
Building the knowledge, practical skills, and organisational readiness needed to use AI safely, ethically, and effectively, without requiring specialist expertise your organisation does not have.
Safety and Ethics at the Core
Embedding responsible AI practices through accessible governance tools and decision-making support that protect both organisations and the communities they serve, before harm occurs, not after.
Burden Reduction
Reducing administrative, operational, and cognitive burden by providing tools that make AI adoption achievable for resource-constrained organisations, not frameworks that create more work than they solve.
Future Readiness and Community Impact
Supporting organisations to anticipate and prepare for emerging AI impacts on communities, including displacement, digital exclusion, and the erosion of human connection in service delivery.
Who this is for
Three groups. Three starting points. One shared outcome.
Not For Humans resources are designed to work across an entire organisation, giving each group what they need to play their part.
Teams and Frontline Staff
Staff, coordinators, case workers, team leaders
Permission and language. Permission to pause without being seen as obstructive. Language to name what you are noticing in AI-assisted work. Clarity about when to escalate. The resources at this layer are built to be picked up in a team meeting this week, with no preparation and no specialist knowledge.
Human Behaviour CardsLeaders and Boards
Directors, managers, boards, risk and governance leads
Line of sight into how AI is actually being used, and confidence that decisions are being made responsibly under pressure. The resources at this layer run from an honest diagnostic through private individual reflection to a committed, owned, dated plan.
Organisational ResourcesFacilitators and Peak Bodies
Consultants, sector infrastructure organisations, capacity builders
If you support multiple organisations, our resources are built to be delivered rather than only used. Facilitator licensing is in development. Register your interest and we will come to you when it is ready.
Facilitator ResourcesAnd the communities you serve. They are not a customer. They are the outcome. Decisions affecting them get made with more thought. Risk is managed before harm occurs. Lived experience stays in the room when AI speaks first.
Self-guided, low cost or free, AI agnostic, and evidence-grounded. The full reasoning is on our About page.
Why now
There is a date on this, and it is closer than it looks.
From 10 December 2026, new transparency obligations under the Privacy Act 1988 commence. Where an organisation has arranged for a computer program to use personal information to make a decision that could reasonably be expected to significantly affect a person’s rights or interests, that must be disclosed in its privacy policy. The regulator’s own examples include decisions affecting a person’s access to a significant service or support.
Intake. Eligibility. Triage. Waitlist prioritisation. That is the description.
Be clear about what this does and does not cover. The obligation is about automated decisions that significantly affect rights or interests. It does not cover AI use generally, and someone drafting a letter in ChatGPT is not in scope. The Privacy Act applies to organisations with turnover above $3 million, and to health service providers at any size, which includes a large part of this sector regardless of turnover.
What the deadline does is force a question: which of our systems make, or substantially assist in making, decisions about the people we serve? Most organisations cannot answer that today. Answering it is an inventory, and the inventory is where this work starts.
This is general information, not legal advice.
OAIC guidance on APP 1 ↗The inventory is where our resources start too. The AI in Use Audit and the AI Register exist for exactly this question, and everything else in the suite builds the capability that makes the answer mean something.
Not For Humans builds AI-capable humans in not-for-profits, NGOs, and community organisations, through behavioural awareness and practice.
Short term, confidence replaces anxiety. Medium term, capability becomes practice. Long term, the sector becomes resilient rather than reactive.
Not For Humans: Because AI’s greatest risk isn’t technical, it’s Human.