People and Culture Leadership in the AI Age: What Matters Now?
This cross-sector session brought senior people and culture leaders together to examine what organisation-wide AI change now demands of their profession. Facilitated by Dr Jan Anderson and Chloe Hawcroft of People Measures, the workshop deliberately set aside the familiar territory of policies, protocols and governance frameworks to focus on the messier human questions sitting behind them — trust, capability, fairness, and how work itself is changing. Through anonymous polls, table discussions and a shared debrief, participants from across the Australian Public Service, private firms and academia compared what they are seeing, where they are getting traction, and where the genuine challenges lie. The recurring theme was that AI adoption only delivers when people engage with it, and that people and culture leaders have a pivotal, and often under-recognised, role to play in making that happen.
Workshop reflections
Overview
- AI adoption is fundamentally a human and culture challenge, not an IT one; technical guardrails, governance and training are necessary but not sufficient, and the assumption that AI "belongs" to IT or a single AI officer leaves the hardest workforce questions unowned.
- Tensions between enthusiasm and reluctance exist in every organisation, with people spread along a spectrum from cautious to impatient; research cited in the session suggested that even where enterprise investment is high, much workplace AI capability goes unused while unsanctioned tool use — including with confidential data — is common.
- AI change is adaptive change: no one yet has all the answers, every advantage carries trade-offs, and progress depends on building enough trust to hold hard conversations safely.
- Access is not adoption — top-down rollouts of licences often produce an early spike in use that then falls away, because people lack permission, purpose and the confidence to experiment safely.
- The entry-level workforce is a pressure point: AI can make junior staff faster, but hollowing out junior roles risks breaking the pipeline that produces the experienced judgement needed to supervise AI-augmented work.
- Hidden costs, shifting pricing models and dependency risks raise questions of data sovereignty and long-term resilience that reach well beyond individual tool choices.
- People and culture leaders need to be at the table with technologists — not as the voice that only asks "but what about the people", but as partners who can balance risk and opportunity, translate between worlds, and steward workforce transformation.
- Strengthening capability across the profession calls for shared language, cross-sector learning and structured peer exchange, learning from sectors that are further along.
Key insights
01
AI adoption is a people problem, not just an IT one
Much of the public conversation about AI readiness centres on instructions, protocols and policies. The harder, more consequential work is human: helping people understand their own accountabilities, managing differing responses to change, and sustaining trust. Treating AI as the responsibility of IT alone — or of a single senior AI role — leaves the workforce and culture dimensions without a clear owner.
02
Literacy is continuous, not a single workshop
A persistent misconception is that AI literacy is achieved by sending people to a one-off course. In practice, capability uplift is ongoing and depends heavily on how managers support continued learning in the flow of work. The question for people leaders is less "have they been trained" and more "how is that learning reinforced day to day".
03
Tensions are universal and worth surfacing
No organisation has everyone on board. Some staff are reluctant; others are frustrated by how slowly things move. These tensions are healthy to surface rather than suppress, and people and culture professionals are well placed to facilitate the psychologically safe, sometimes difficult, conversations that adaptive change requires.
04
Ethical principles are a practical conversation tool
Framing adoption around benefit to people and society, fairness and inclusion, contestability, and workforce planning gives leaders a structured way to ask better questions: Will this genuinely benefit our people? Who might be excluded? How do staff contest decisions that affect their roles? Inclusion cuts both ways — assumptions that particular groups, such as mature-age workers, are uninterested in technology can quietly leave them out.
05
Access does not equal adoption
Giving everyone a licence is not the same as enabling real use. Rollouts frequently see an initial surge followed by a sharp drop-off, because people do not know what to use the tool for or how to use it safely and responsibly. Sustained adoption needs explicit permission to experiment and a "safe to fail" environment — distinct from "fail safe" — where mistakes become learning rather than something to drive to zero. The tolerance for experimentation also differs sharply between internal-facing work and live service delivery.
06
Protect the entry-level pipeline
AI lets more junior staff produce sophisticated work faster, which tempts organisations to thin out entry-level roles. But experienced judgement is built, not bought; cutting the bottom rungs risks removing the very people who later supervise and direct AI-augmented work well. This is a tension between organisational productivity (fewer people, more tools) and the wider productivity of having a healthy talent pipeline — and some sectors that cut entry-level programs have since reversed course.
07
Train behaviours, not just tools
Training tends to focus on a specific product rather than the durable behaviours and judgement that apply across any tool. It also under-counts the rework involved when AI output still needs a human layer added over the top. Designing for the underlying capabilities, not the current interface, makes the investment more resilient as tools change.
08
Watch the hidden costs and dependency risk
Shifting pricing models — particularly usage- or token-based ones — mean staff can incur real costs without realising it, and heavy reliance on a tool can leave an organisation exposed if prices change later. Alongside this sit harder questions of data sovereignty: who holds the data, who makes decisions about it, and how much access tools quietly accumulate to sensitive material such as email and notes.
09
Lead with role clarity and honesty about uncertainty
Getting role definitions, work-level standards and architecture right gives people a clear sense of the pathway through change. Equally important is transparency: being open about what is known, what is still being worked out, and where the organisation is genuinely experimenting tends to make people more comfortable, not less.
10
People and culture belongs at the decision table
People and culture functions have long led capability uplift and difficult conversations, which places them at the centre of this moment. The risk is being cast as the people who only push back on behalf of staff; the opportunity is to help balance risks and opportunities, translate between technical and human framings, and own the workforce transformation work — identifying which roles are affected and how to support, retrain or transition people — that no other function can.
11
Normalisation is the destination
A useful reframe is that AI will eventually become "just another tool", much as email did. That invites two questions: what has reliably worked in past change efforts that still applies, and what is genuinely new and needs specific treatment. Holding both prevents AI from being treated as an exotic problem sitting outside normal work.
12
Build capability collectively across the profession
The strongest suggestions for strengthening the profession were collective: more cross-sector forums, peer learning circles and action-learning sets, and deliberately learning from sectors further along the curve. A shared, plain language — translating "tech-speak" into what change means for ordinary staff — emerged as a distinctive contribution people and culture leaders can make to the broader public-sector conversation.
AI CoLab evaluation
Summary from discussion
Lowering barriers to safe experimentation
- The session itself modelled low-stakes participation, using anonymous polling, table brainstorms and a shared debrief so participants could contribute candidly without exposure.
- A central theme was creating "safe to fail" conditions inside organisations — explicit permission to experiment, space to reflect and adapt, and treating mistakes as learning — as the precondition for adoption to move beyond a brief initial spike.
- The facilitators offered a no-cost, cross-sector peer learning circle for chief people officers as a concrete, low-barrier way to keep experimenting and comparing notes after the session.
Uncovering barriers to practical collaboration
- Real blockers surfaced repeatedly: fragmented, project-by-project rollouts without an overarching people strategy; training pitched at tools rather than behaviours; and uneven adoption that cannot be reduced to age or seniority.
- Structural and economic frictions were prominent — shifting and usage-based pricing, unanticipated costs, vendor economics and dependency risk, and data-sovereignty concerns including how much sensitive data tools can quietly access.
- The hollowing-out of entry-level roles was identified as a systemic risk to the future supply of experienced judgement, set against a tension between organisational and societal productivity.
Building links between technical experts and reformers
- A consistent message was that people and culture leaders must be in the room with technologists, so that workforce and culture decisions are not made by technical leaders without people expertise.
- Participants framed their distinctive contribution as language and framing — surfacing the right questions and translating technical concepts for the wider workforce — and as the connective tissue between strategy, technology and the lived experience of staff.
- The proposed learning circle and cross-sector networks were positioned as practical mechanisms to connect practitioners and feed shared learning back into the broader profession.
Range and diversity of participants
- The room spanned multiple Australian Public Service agencies, private-sector firms, corporate and finance functions, and academia and research, reflecting a genuinely cross-sector mix.
- Roles skewed senior — chief people officers, heads of workforce capability, and chief operating officers — bringing both strategic and operational vantage points to the discussion.
- Participants also brought a range of generational perspectives and personal stances on AI, which directly informed the conversation about meeting different cohorts where they are.
About the facilitator
The session was facilitated by People Measures, a leadership-development consultancy that assesses and develops leadership performance and culture and increasingly works at the intersection of people and AI. Dr Jan Anderson is Director Innovation, with chief AI officer-equivalent responsibilities, leading ethical AI adoption across the business; Chloe Hawcroft is CEO, with extensive people and culture leadership experience across the public and private sectors. Further detail about the session is available on the AI CoLab event listing.