Every people function is having the same conversation right now. AI in HR has moved from pilot project to boardroom expectation in the space of about eighteen months, and nobody wants to be the function that got left behind, or the function that adopted it recklessly and broke something that mattered.
That is the dilemma. Not whether to engage with AI, that decision has already been made for us. The real question about AI in HR is how to adopt it without hollowing out the things that make an organisation worth working for: capability, trust, fairness, and the conditions for people to actually do good work.
The scale of change is not in dispute
The numbers on AI-driven labour market change are large enough that they are easy to wave away as abstract. The World Economic Forum’s Future of Jobs Report 2025, drawn from a survey of more than 1,000 employers across 55 economies, puts a figure on it: 170 million jobs created and 92 million displaced globally by 2030, a net gain, but a churn equivalent to 22 percent of all formal jobs in the dataset. Nearly 40 percent of core skills are expected to change in that window.
Those are global, macro numbers, and it is tempting to treat them as someone else’s problem. They are not. Every one of those transitions happens inside an organisation, managed by a people function, and the manner of that management is what determines whether the transition builds trust or destroys it.
Adoption is outpacing redesign
Deloitte’s 2026 Global Human Capital Trends report, based on a survey of over 9,000 leaders across 89 countries, makes the sharpest version of this point. Organisations are moving quickly to adopt AI, but far fewer are redesigning work, leadership and culture to go with it. The report finds that two-thirds of C-suite leaders believe traditional functions like HR, finance, IT and legal need to change, yet only a small fraction believe they are actually making progress on that change.
That gap between adoption and redesign is where the dilemma in AI in HR actually lives. Buying a tool and mandating a rollout is straightforward. Redesigning a role so it still builds capability, still offers judgement and craft, and still gives an employee a reason to trust the process, is much harder, and it is the part most organisations skip.
The cost nobody puts on the business case
CIPD research published this year has started naming something practitioners have felt for a while: a form of mental fatigue that comes specifically from overseeing AI output rather than doing the work directly. It differs from ordinary burnout. It is driven by constantly checking, correcting and second-guessing a tool rather than exercising a skill.
That matters for how AI in HR gets designed at task level. If a role is restructured so a person spends their day supervising an algorithm instead of practising a craft, the job has not automatically improved. It may have got worse in a way that will not surface until retention drops or engagement scores dip, well after the business case has been signed off.
There is an equity dimension too. Entry-level roles tend to be more exposed to AI displacement than senior ones, largely because so much of what juniors traditionally learn on the job is the routine work AI now does faster. Automate that work without redesigning early-career development alongside it, and organisations are not just cutting cost, they are dismantling the pipeline that produces their future senior talent. CIPD is right to frame this as a workforce planning problem rather than a change management footnote.
Where the real skill in AI in HR sits
CIPD’s own research, based on a survey of over 1,300 senior leaders and HR professionals conducted in early 2026, adds a useful correction to the standard narrative. The constraint on responsible AI adoption is rarely a skills gap in the traditional sense. HR and leadership teams already hold the core capabilities: workforce planning, job design, employee engagement. The gap is in applying those capabilities to a form of change that does not behave like the change most organisations are built to handle.
The people professionals who get AI in HR right will be asking a different set of questions before deployment, not after it. Which tasks are we automating, and what was a junior colleague learning by doing them? What happens to trust and fairness when AI drafts employee-facing correspondence and someone still has to catch every inaccurate or irrelevant point it introduces? What does good work look like for a role once the parts requiring judgement have been separated from the parts requiring oversight and correction?
There is a useful shorthand doing the rounds in workforce planning circles this year: buy, build, borrow or bot. Deloitte’s 2026 report extends it further, adding options like blending human and machine effort rather than treating them as a straight substitution. It is a decent frame, but it only works if the AI option is interrogated with the same rigour as the other three.
The dilemma is a design problem, not a technology problem
The organisations that get through this well will not be the ones with the most sophisticated AI stack. They will be the ones treating AI in HR as an organisational design question from the outset: what should a role look like once AI has taken on part of it, who is accountable for the parts that remain, and how do we protect the development pathways that used to run through the tasks we have just automated.
That is squarely HR and OD territory. It always has been. The AI dilemma is not really about AI. It is the same question the profession has been answering for decades, about how work gets designed, how capability gets built, and how fairness gets protected when the ground shifts quickly. The tools are new. The job is familiar.
Where to start
If your organisation is further into AI adoption than it is into AI-ready workforce design, that gap is worth naming before it widens. The Change Readiness Tool on Avant People is a practical starting point for surfacing exactly where that gap sits in your own organisation. If you want to talk through what redesign looks like for your teams specifically, get in touch, that conversation is where the selective consultancy work starts.
Sources
- World Economic Forum, Future of Jobs Report 2025
- Deloitte, 2026 Global Human Capital Trends
- CIPD, Futureproofing your skills: AI is changing how HR skills are applied (2026)
- CIPD, Using AI to raise productivity without weakening capability and trust (2026)
- CIPD Festival of Work 2026, as reported by HRZone and People Management