Habit
The individual
The person at the keyboard. The questions only they can answer about their own AI use and the discipline that turns into a daily practice.
AI Accountability Map
HUMAN. The five levels every Australian leadership team must hold themselves to.
HUMAN is Tracy Sheen’s proprietary accountability map for AI governance in Australian organisations. Five levels in stack. Habit (the individual). Unit (the team). Macro (the organisation). Audience (clients and suppliers). Neighbourhood (community). Each level holds itself to the same four GIST questions: what protects us, why are we using AI, what is the plan, what must we learn. The result is governance that holds from the boardroom to the frontline.
Why HUMAN exists
Most AI policies live at one level. Then everyone goes back to work.
The pattern repeats. The board signs off on an AI policy. IT rolls out a tool. Communications drafts an internal note. Done. Filed. Move on.
Meanwhile the actual accountability lives at four other levels nobody is looking at. The individual at the keyboard. The team trying to coordinate. The clients on the receiving end of the work. The community the organisation operates in.
That is how AI policies become paperwork. Signed at one level. Ignored at the four levels where it counts.
HUMAN names the five levels and asks the same four GIST questions at each one. Twenty conversations the leadership team needs to be having. Not five.
AI accountability has five levels. Stop having a one-level conversation.
48%
of Australian employees use AI in ways that contravene company policies. Habit-level accountability gaps are real.
36%
of organisations globally have a formal AI governance policy. Macro-level work is still missing in nearly two-thirds of organisations.
83%
of Australians say they would trust AI more when international standards and responsible governance are in place. Audience and Neighbourhood levels matter.
The five levels
From the keyboard to the community.
HUMAN stacks accountability across five stakeholder levels. Each level is a different conversation. Each level demands different evidence.
Unit
The team
The team or department. Coordination, not one-by-one decisions. So that one person’s AI use does not create risk for the others.
Macro
The organisation
The org-wide layer. Policy, compliance, capability uplift, board-level oversight. Where most AI governance work stops today.
Audience
Clients and suppliers
The people on the other side of the work. Trust, transparency, the way AI use shows up in client outcomes and supplier relationships.
Neighbourhood
The community
The community the organisation operates in. Ethical responsibility, AI literacy contribution, the broader question of what good looks like.
The matrix
Twenty questions. The complete accountability map.
Each HUMAN level (rows) crossed with each GIST pillar (columns). Twenty cells. Twenty conversations every leadership team needs to be having about AI.
Hover a cell to highlight its row and column.Click a cell to expand the question.
HUMAN ↓
GIST →
GIST →
Guardrails
Intent
Strategy
Practical Training
Habit
H · GWhat guardrails do I personally need to use AI safely?
H · IWhy am I using AI? What am I trying to achieve?
H · SWhat is my personal AI plan? What do I use, when and for what?
H · TWhat do I need to learn first to use AI well and safely?
Unit
U · GWhat do we agree on so one person’s AI use does not create risk for the others?
U · IAre we aligned on what AI is for in our team?
U · SHow do we coordinate AI use, so it compounds rather than conflicts?
U · TWhat does the team need to learn together so nobody is left behind?
Macro
M · GWhat policies, compliance requirements and risk frameworks need to be in place?
M · IWhat is our organisational AI intent? Shared purpose, or everyone doing their own thing?
M · SWhat is the org-wide rollout plan? Phased, modular, measurable?
M · TWhat is the capability uplift program? Who owns it and how does it stick?
Audience
A · GWhat guardrails protect client data and supplier relationships?
A · IHow does our AI use affect the outcomes we deliver to clients?
A · SHow do we communicate our AI strategy in a way that builds trust externally?
A · TWhat do clients and suppliers need to understand about how we use AI?
Neighbourhood
N · GWhat is our ethical responsibility to the broader community in how we use AI?
N · IWhat is our intent as an organisation toward the communities we operate in?
N · SHow does our AI strategy contribute to or detract from community wellbeing?
N · TWhat is our responsibility to build AI literacy in the communities we serve?
Macro · Strategy
What is the org-wide rollout plan? Phased, modular, measurable?
Macro (Organisation)
The org-wide layer. Policy, compliance, capability uplift, board-level oversight. Where most AI governance work stops today.
Strategy
Strategy is the GIST pillar that sequences rollout against business goals. Built using the intelligence already in the room.
HabitIndividual
UnitTeam
MacroOrganisation
AudienceClients & Suppliers
NeighbourhoodCommunity
Macro · Strategy
What is the org-wide rollout plan? Phased, modular, measurable?
Macro (Organisation)
The org-wide layer. Policy, compliance, capability uplift, board-level oversight. Where most AI governance work stops today.
Strategy
Strategy is the GIST pillar that sequences rollout against business goals. Built using the intelligence already in the room.
Australian standards
Accountability mapped to the standards your board already knows.
HUMAN does not replace existing governance. It maps the stakeholders so the existing standards land at every level they need to.
Voluntary AI Safety Standard ↗
The Standard’s ten guardrails apply across all five HUMAN levels. HUMAN names which stakeholders own each guardrail conversation.
Australian AI Ethics Framework ↗
Accountability is one of the eight ethics principles. HUMAN operationalises it across the five-level stack so accountability is not just an org-wide claim.
ISO/IEC 42001:2023 ↗
The standard’s clauses on organisational context (Macro), leadership (Macro) and stakeholder needs (Audience and Neighbourhood) map cleanly onto HUMAN.
Privacy Act 1988 + APPs ↗
Data handling responsibilities cascade across all five HUMAN levels. The Habit and Unit levels are where most APP breaches actually happen.
In practice
Where HUMAN shows up in the work.
The matrix lives in three places across the work Tracy delivers. It is a working document, not a poster.
Boardroom briefings
The matrix turns into the agenda. Each cell is one boardroom conversation. The leadership team works through the cells in priority order and identifies who owns what.
Discovery and Blueprint phases
HUMAN is the stakeholder map for any GIST engagement. Used to make sure no level is missing from the AI plan. The matrix becomes a working document the team revisits as rollout progresses.
Conference keynotes
The audience leaves with the matrix to apply at their own organisation. The five-level framing shows them what their current AI policy is missing without naming any one company.
Run the conversation yourself
The HUMAN Leadership Facilitation Guide.
A working guide for leadership teams ready to walk the matrix themselves. Twelve pages. Twenty questions. The exact prompts and signals to listen for at each level.
How it fits
HUMAN is one of three sister frameworks.
GIST is the methodology. HUMAN keeps the methodology honest at every level of accountability. REIMAGINE rolls it out as a change arc.
GIST Framework
The four-pillar methodology.
Guardrails, Intent, Strategy, Practical Training. Each pillar protects the next. The questions HUMAN asks at every level.
Read more →REIMAGINE FrameworkThe nine-step change arc.
From Relationships through Evolve. Built on the research in AI & U: REIMAGINE Business. Names the stakeholders REIMAGINE has to bring with it.
Read more →Common questions
HUMAN. The questions that come up most.
What is the HUMAN framework?
HUMAN is Tracy Sheen’s proprietary five-level accountability map for AI in Australian organisations. Habit (individual), Unit (team), Macro (organisation), Audience (clients and suppliers), Neighbourhood (community).
Each level holds itself to the same four GIST questions. The matrix is twenty cells in total.
Is HUMAN a diagnostic tool?
No. HUMAN is an accountability map, not a diagnostic tool. It tells leadership teams which stakeholders need to be in the AI governance conversation. It does not score organisations or rank maturity.
The distinction matters because diagnostics produce reports. Accountability maps produce action.
How does HUMAN relate to GIST?
GIST is the methodology. HUMAN is the layer that keeps GIST honest at every stakeholder level.
The four GIST questions get asked at all five HUMAN levels. The matrix shows what changes at each combination, twenty cells in total.
Where do leadership teams typically use HUMAN?
Most engagements use HUMAN in the Discovery and Blueprint phases of corporate strategy work. The leadership team uses the matrix to identify which stakeholders the AI plan must account for.
The matrix becomes a working document the team revisits as the rollout progresses.
Can HUMAN be applied to small business?
Yes. The matrix works the same way at any organisation size.
For small businesses and solopreneurs the Habit and Unit levels carry more weight, the Macro level is lighter and Audience and Neighbourhood become highly visible. Small business owners apply it through the Practical AI Collective community.