TL;DR: AI won't eliminate most UK jobs by 2030 — but it will reshape them significantly. This guide helps HR directors understand which roles are most exposed, what reskilling and legal obligations apply, and how to lead workforce transformation ethically and strategically.
Introduction: The Question HR Directors Actually Need to Answer
The headlines are hard to ignore. "300 million jobs at risk from AI." "Half the workforce automated by 2030." If you've been in an executive meeting recently, someone has almost certainly dropped one of these figures on the table and looked to HR for an answer.
Here's the thing: that's not quite the right question. The more useful question — the one HR directors should actually be losing sleep over — is this: What is HR's responsibility to prepare the workforce for what's coming?
The scale of change is real. According to the Tony Blair Institute for Global Change and BCG (2026), 70% of UK jobs have high AI exposure, meaning more than half of their tasks could be automated. But that figure needs immediate context: BCG's own analysis distinguishes sharply between jobs being reshaped and jobs being replaced. Only 10–15% of UK jobs face outright elimination by 2030. The remaining 50–55% will change significantly — but they won't disappear.
That distinction matters enormously for how HR responds. This guide covers which UK roles are most exposed, what the law requires of you when AI drives restructuring, and how to lead reskilling and workforce transformation ethically. It's written for HR directors who are caught between genuine business pressure to adopt AI and a real duty of care to the people it affects.
Which UK Jobs Are Most Exposed to AI by 2030?
Not all jobs face the same level of risk, and understanding the difference between "reshaped" and "replaced" is the starting point for any serious workforce planning conversation.
BCG's 2026 analysis is clear: 50–55% of UK jobs will be significantly reshaped by AI, but only 10–15% face outright elimination. The roles most at risk share a common characteristic — they are dominated by repetitive, rules-based tasks that require relatively low levels of human judgment.
The data on specific roles is striking:
| Role | Automation Risk | UK Jobs Affected | Median Wage |
|---|---|---|---|
| Bookkeepers / Payroll Clerks | 94% | 249,000 | £25,889 |
| Bank / Post Office Clerks | 87% | 82,000 | — |
| Pensions / Insurance Clerks | 87% | 30,000 | — |
| Customer Service Representatives | 76% | 365,000 | — |
Source: Logic4training / BCG / Tony Blair Institute, 2026; LiveCareer, 2026 (ILO/ONS data)
Customer service representatives represent the largest volume — 365,000 UK jobs with 76% automation risk. For HR directors in retail, financial services, or any customer-facing sector, that's not an abstract statistic. It's a significant portion of your workforce.
But consider the M&S warehouse example, which illustrates the nuance well. When M&S introduced AI-powered systems in its distribution operations, accident rates fell by 80%. The workforce wasn't eliminated — the work changed. Employees shifted from purely manual tasks to roles involving oversight, exception handling, and quality control. That transition required active HR management. It didn't happen automatically.
The global picture offers some reassurance. The World Economic Forum's Future of Jobs Report (2025) projects 92 million global job displacements by 2030 — but also 170 million new jobs created, a net positive of 78 million roles. The challenge isn't the destination; it's managing the transition. And that transition is squarely an HR responsibility.
One figure that should concern UK HR directors specifically: fewer than 50% of UK workers have been offered any AI training, compared to Germany and the United States (Logic4training, citing BCG/Tony Blair Institute, 2026). That gap isn't a government problem to solve in isolation — it's a workforce planning failure that HR leaders can directly address.
Which Jobs Are Safest? (And What Bill Gates Actually Said)
If you've searched "which 3 jobs will survive AI" recently, you're not alone — it's one of the most searched questions on this topic. The honest answer is that no job is entirely immune, but some are far more resilient than others.
Bill Gates, in 2024 interviews, identified teachers, therapists, and inventors as the roles least vulnerable to AI. The common thread is clear: all three require human empathy, creative originality, or ethical judgment that AI cannot reliably replicate. A therapist's ability to read unspoken distress, a teacher's capacity to adapt to a struggling student in real time, an inventor's lateral thinking — these are genuinely difficult to automate.
Google's AI Overview takes a slightly different lens, pointing to AI programmers, biological scientists, and energy sector specialists as roles that will survive and grow. Both perspectives are valid — they're just measuring different things. Gates is identifying roles where human connection is the core value; Google's answer identifies roles where human expertise remains irreplaceable in technically complex domains.
In a UK context, there's a third category worth highlighting: skilled tradespeople. Plumbers, electricians, and gas engineers face fewer than 10% automation risk, according to Logic4training. Physical dexterity, on-site problem-solving, and regulatory compliance requirements (Gas Safe registration, for instance) create natural barriers to automation that no large language model can overcome.
The pattern across all genuinely resilient roles is consistent: human judgment, physical presence, emotional intelligence, or creative originality. If a role requires any of these in significant measure, it is substantially protected. If it requires none of them, it is exposed.
So what does this mean for HR's own role?
Will AI Replace HR Professionals?
The direct answer: no, not in any meaningful sense — but with an important caveat.
CIPD research (2025) finds that 85% of HR tasks are resilient to automation. That's a reassuring headline, but the nature of those resilient tasks is shifting in ways HR teams need to understand. For a deeper look at how AI is changing the day-to-day reality for HR teams, the picture is more nuanced than a simple replacement narrative suggests.
What AI will handle — and is already handling in many organisations — includes repetitive policy queries, interview scheduling, initial CV screening, benefits administration, and compliance reminders. These are tasks that consume significant HR bandwidth but require relatively little human judgment. Tools like Aura handle exactly this kind of volume: the "what does this restructuring mean for my role?" and "am I entitled to redundancy pay?" questions that flood HR inboxes during periods of change, freeing HR professionals to focus on the complex conversations that require genuine human judgment.
What humans must own is everything that matters most: complex employee relations, redundancy consultations, culture-building, ethical decision-making, and strategic workforce planning. These are not tasks AI can perform — they are tasks AI can support, but the accountability and judgment must remain human.
New roles are also emerging that didn't exist five years ago: AI ethics lead, workforce transition manager, prompt engineer for HR systems. These are HR-adjacent roles that require the kind of organisational understanding and people expertise that HR professionals already possess.
The real risk for HR isn't replacement. It's irrelevance. HR teams that don't engage with AI — that treat it as an IT problem or a passing trend — will be outpaced by those that do. The profession's resilience depends on HR actively shaping how AI is deployed, not passively observing it.
HR's Legal Obligations When AI Drives Redundancy
This is the section most articles on AI and jobs completely skip — and it's arguably the most important for HR directors to get right.
AI-driven restructuring does not exist in a legal vacuum. The same employment law framework that applies to any redundancy programme applies when the driver is automation. Critically, it applies whether or not your organisation acknowledges AI as the cause.
Under the Employment Rights Act 1996 and TULRCA 1992, collective consultation is required where 20 or more redundancies are proposed within a 90-day period. That means a minimum of 45 days' notice before the first dismissal takes effect. There are no AI exemptions. If your automation programme results in 20+ roles becoming redundant, the consultation clock starts ticking from the moment the decision is made — not when you announce it to employees.
The Equality Act 2010 introduces a specific risk that many HR teams are not yet managing. If AI tools are used to identify which roles or individuals are selected for redundancy, those tools must be audited for algorithmic bias. The Equality and Human Rights Commission (EHRC) reported a 25% rise in algorithmic discrimination claims in 2025. If an AI system disproportionately flags roles held predominantly by women, older workers, or ethnic minorities for elimination, the legal exposure is significant — and "the algorithm decided" is not a defence.
The shadow AI risk in the workplace compounds this problem. Where employees or managers are using unapproved AI tools to inform decisions — including decisions about workforce structure — HR may be unaware of the bias risks being introduced.
TUPE 2006 is also relevant where AI automation accompanies outsourcing decisions, a combination that is increasingly common in back-office functions.
Looking ahead, the AI Regulation Bill (currently in 2025 consultation) proposes transparency requirements for high-risk AI applications in recruitment and redundancy decisions, with potential fines of up to 4% of global turnover for non-compliance. This is not yet law, but the direction of travel is clear.
The financial stakes of getting this wrong are concrete. Employment Tribunal cases for technology-driven restructures rose 15% in 2025 (Ministry of Justice, 2025). The average SME redundancy settlement runs between £15,000 and £30,000 (ACAS, 2025) — before legal costs. Statutory redundancy pay is calculated at a minimum of 1.5 weeks' pay per year of service, capped at £700 per week (2026 rate).
None of this is unmanageable. But it requires HR to lead proactively, not reactively.
Building Your Reskilling Strategy: A Practical Framework
There is currently no statutory reskilling mandate in UK law. However, employers have an implied duty of care under the Health and Safety at Work Act 1974 to ensure employees are equipped for their roles — and CIPD guidance is clear that proactive reskilling is both good practice and a meaningful risk mitigation tool.
Here's a practical five-step framework for HR directors building their reskilling response.
Step 1 — Audit your exposure. Map your workforce against ONS Standard Occupational Classification (SOC) codes and cross-reference with the automation risk data. Identify the 20% of your workforce in high-exposure categories. This is your planning baseline.
Step 2 — Scenario plan. Use BCG's 10–15% elimination rate as your conservative planning assumption for three-year forecasts. Model what your workforce looks like in 2027 under different automation adoption scenarios. This gives you a defensible basis for board-level conversations.
Step 3 — Prioritise reskilling investment. CIPD-approved AI literacy courses via FutureLearn cost between £500 and £2,000 per employee. McKinsey's 2025 UK benchmark is 20% of the workforce trained in AI-relevant skills by 2027. If fewer than 50% of UK workers have been offered any AI training, getting ahead of that benchmark is a genuine competitive advantage.
Step 4 — Use the Apprenticeship Levy. This is one of the most underused tools available to UK employers for AI upskilling. Digital and data apprenticeships are eligible for Levy funding, and many organisations are sitting on unspent Levy balances. If your CFO is asking how to fund reskilling, this is part of the answer. For further resources, CIPD's 2025 AI HR Toolkit at cipd.org provides practical guidance on building an AI-ready workforce.
Step 5 — Communicate early and honestly. Employees who understand the plan are significantly less likely to leave. When 53% of IT leaders expect workforce shrinkage from AI adoption, the rumour mill starts early. Transparency about what is changing, what support is available, and what the timeline looks like is not just good ethics — it's good retention strategy.
The employee questions don't stop during a restructure — they multiply. "What does this mean for my role?" "Will I be made redundant?" "What training will I get?" Having a system that can answer policy questions accurately and instantly, at any hour, means your HR team can focus on the consultations that actually matter. That's exactly what Aura is designed to do during periods of workforce change: absorb the volume of repetitive queries so HR can focus on the complex, human conversations.
One budget reality to acknowledge: CIPD (2025) finds most UK SMEs allocate less than 5% of their budget to training. If you're making the case to your CFO for reskilling investment, frame it against the alternative — the average SME redundancy settlement of £15,000–£30,000, multiplied by the number of roles at risk. Reskilling is almost always cheaper than redundancy, and it preserves institutional knowledge.
The Ethical Dimension: Leading AI-Driven Change Responsibly
Legal compliance is the floor, not the ceiling. The hardest part of leading AI-driven workforce change isn't the paperwork — it's the human cost.
Ethical redundancy practice means offering voluntary reskilling before compulsory redundancy, running genuine 30 or 45-day consultations rather than tick-box exercises, and ensuring that employees understand not just what is happening but why. The distinction between a consultation that informs and one that genuinely listens is one employees will notice — and remember.
One practice to avoid categorically: using AI tools to select who gets made redundant without human review. This is both legally risky, as noted above, and ethically indefensible. Algorithmic outputs must be reviewed by a human who can assess context, apply judgment, and take accountability for the decision. The ILO's benchmark target is for 13.4% of the workforce to be augmented rather than displaced by AI — tracking your own ratio via retention KPIs post-implementation is a useful internal measure.
In unionised workplaces, involve employee representatives and works councils early in AI planning — not as a legal formality, but as a genuine input into how change is managed. Employees who feel involved in the process are more likely to support it.
The reputational dimension is real and lasting. How your organisation handles AI-driven change will define your employer brand for years. Employees talk. Glassdoor reviews are permanent. The organisations that will attract talent in 2028 and beyond are the ones that handled the transition of 2025 and 2026 with integrity.
Understanding how to build an AI use policy for your workforce is a practical starting point — it creates the governance framework within which ethical AI decisions can be made consistently, rather than case by case.
The reframe that matters: HR's role in the AI era isn't to manage decline. It's to lead transformation. That's a fundamentally different posture — and it's one that the profession is well placed to take.
Conclusion: HR's Moment to Lead
AI will reshape far more UK jobs than it eliminates by 2030. BCG's data is clear on this: 50–55% of roles will change significantly, but only 10–15% face outright elimination. The transition, however, requires active HR leadership — it will not manage itself.
Three things HR must do now: audit your workforce's exposure using the available data, build a reskilling plan that makes the business case to your CFO, and lead ethical change with transparency and genuine consultation. These are not optional extras. They are the core of what HR is for in this moment.
HR professionals themselves are among the least replaceable people in any organisation — but only if they engage with AI rather than avoid it. The profession's resilience is earned, not guaranteed.
If your team is facing a period of workforce change and the employee questions are already multiplying, see how Aura helps HR teams manage the volume — so your people can focus on the strategic, human work that only they can do.
