David Autor, an economist at MIT and one of the most cited researchers on technology's impact on labour markets, has spent decades tracking what happens to work when automation arrives. His conclusion: the jobs that disappear are always more visible than the jobs that appear.

The jobs AI is eliminating appear in headlines. But the jobs AI is creating do not yet have established degree programmes or standardised job descriptions. And most HR teams are trying to hire for them using frameworks built for jobs that do not exist anymore. That gap between the roles AI is creating and the HR capability to find people for them is a major workforce challenge of 2026.

AI job creation is in full swing

According to the World Economic Forum's Future of Jobs Report 2025, AI and automation are expected to create 69 million new jobs globally by 2027, while displacing 83 million. The new jobs are not direct replacements for the old ones. They are completely new roles that require combinations of skills no single traditional career path has produced. They sit at the intersection of human judgment and machine capability, and they are already open, in organisations struggling to fill them. Here are five of the most important ones.

---

Role 1: AI behaviour analyst

AI systems deployed within organisations frequently produce unintended outputs: flagging the wrong candidates, misidentifying engagement risk, or giving more weight to the wrong metric. An AI Behaviour Analyst observes AI outputs over time, identifies where the system's behaviour diverges from the organisation's expectations, documents those divergences, and works with technical teams and vendors to correct them.

The problem is that this role does not exist in the larger job market yet — no established pipeline, no job board postings, no university courses. The solution is to hire from adjacent fields with transferable skills: behavioural psychologists who understand how patterns of behaviour deviate from norms, UX researchers trained to spot where a system diverges from user expectations, fraud analysts who watch for signs a system is being gamed, and clinical auditors who assess whether diagnostic tools perform as intended across diverse populations. Look for candidates who have spent their careers asking: why is this system behaving this way, and what does it mean?

Role 2: Human-AI collaboration designer

The AI screening tool is live, but the recruiter does not know when to review its shortlist. The AI engagement tool is live, but the manager does not know how to interpret its signals. The result is AI that is technically operational and practically ignored, because nobody has designed how humans and AI are supposed to work together in specific workflows.

A Human-AI Collaboration Designer builds the workflows, protocols, and decision frameworks governing how people interact with AI tools day-to-day. They define handoff points between automated and human decision-making, build guidance that tells a manager exactly what to do when the AI flags a signal, and measure whether the human-AI workflow outperforms either alone. Source from instructional designers, business process analysts, and human factors engineers from aviation or healthcare, who are specifically trained in designing safe, effective human interaction with automated systems.

Role 3: Workforce transition specialist

When an AI tool automates a significant portion of a role, the employee has three possible futures: their role is eliminated, restructured around the tasks AI cannot perform, or moved into a new role that did not exist before. Each requires a different HR response. A Workforce Transition Specialist manages that complexity at an individual level — mapping transferable skills against new roles, designing personalised transition pathways, and supporting employees through change at a human level. They also work upstream, identifying which roles are most likely to be affected by the next wave of AI adoption so transitions can be planned proactively.

Hire from careers that combined analytical skill assessment with genuine human empathy in high-stakes situations: career coaches experienced in industry transitions, occupational psychologists, social workers or counsellors with commercial acumen, and vocational rehabilitation specialists who are experts at matching existing capability to new opportunity under difficult conditions.

Role 4: AI ethics and compliance officer (HR domain)

AI tools deployed in HR make or inform decisions about people's careers, compensation, wellbeing, and sense of belonging — decisions carrying legal, ethical, and human risk. The AI Ethics and Compliance Officer sits at the intersection of employment law, people strategy, data ethics, and AI governance. They build the frameworks, run the audits, advise the CHRO, and represent the organisation in conversations with regulators, employees, and works councils about how AI is used in the people function.

Look at employment lawyers with a specific interest in technology and data, data protection officers from people-intensive sectors, and HR business partners who have been close to AI deployments and understand the governance gaps — they bring the people context that pure legal or technology backgrounds often lack.

Role 5: People data translator

Most organisations have more people data than ever. AI tools generate workforce insights, engagement signals, performance patterns, attrition predictions, and compensation anomalies at a volume no previous generation of HR tools could produce. The People Data Translator takes this complex, AI-generated workforce data and translates it into specific, actionable recommendations for HR leaders, managers, and executives.

The best hunting ground is science communication — professionals who make complex research accessible to non-specialists. Look at health informatics specialists who translate clinical data into guidance for practitioners, financial analysts who turn quantitative models into recommendations for non-technical clients, and journalists from data-heavy beats who can make numbers tell a story.