The AI Jobs Report: Which Careers Are Growing, Which Are Declining, and What the Data Actually Shows

Every few months brings a new study either catastrophising about AI job displacement or dismissing the concern entirely. Both camps cherry-pick data to support predetermined conclusions. Here is a more careful look at what the labour market evidence actually shows as of 2026, what it suggests for different career paths, and how individuals can use this information strategically.

What the Data Shows

Tasks Are Automating, Not Jobs

The most consistent finding across labour market research is that AI is automating specific tasks within jobs, not eliminating jobs wholesale. A 2025 McKinsey analysis found that 60-70% of occupations have at least 30% of their tasks technically automatable with current AI — but very few jobs consist entirely of automatable tasks.

The practical implication: most workers will find their jobs change significantly rather than disappear. The tasks that remain will tend to be the ones that require judgment, interpersonal skill, or physical presence.

Where Job Postings Are Growing

LinkedIn’s 2026 Jobs on the Rise report shows the fastest-growing roles include:

• AI/ML Engineer — demand up 147% since 2022
• Data Governance Specialist — up 89% (companies need people to manage AI systems responsibly)
• AI Prompt Engineer — emerged as a job category in 2023; stabilising as a secondary skill rather than standalone role
• Cybersecurity Analyst — up 63% (AI creates new attack surfaces that need defending)
• Climate Technology Specialist — up 128% (AI is accelerating climate modelling and clean energy)

Where Job Postings Are Declining

• Data entry and routine clerical work — down 41% since 2022
• Basic coding (junior CRUD development) — down 23% as AI handles more boilerplate code
• Paralegal document review — down 31% in large law firms with AI contract analysis tools
• Junior graphic design (production work) — down 18% as AI image generation handles volume requests
• Telephone customer service — down 29% as AI handles tier-1 support

The Skills That Are Growing in Value

Across multiple labour market analyses, a consistent pattern emerges. The skills most resistant to automation fall into four categories:

1. AI management and oversight — knowing how to work with AI tools, evaluate their outputs, and correct their errors
2. Complex problem-solving with ambiguous constraints — situations where the right approach is not clear and requires genuine judgment
3. Interpersonal and emotional intelligence — sales, therapy, leadership, teaching, negotiation
4. Creative integration — not just generating creative output but deciding what is worth making and why

The “AI Complement” Effect

The most important and underreported finding in AI labour research is what economists call the “AI complement” effect: AI increases the productivity of highly skilled workers more than it increases the productivity of less-skilled workers in the same field.

A study of GitHub Copilot found that experienced developers became significantly more productive with AI assistance, while junior developers saw smaller gains and sometimes became more error-prone (over-trusting AI output they lacked the experience to evaluate). The same pattern appears in legal work, medical diagnosis, and financial analysis.

The implication: AI widens the skill gap. Investing in genuine expertise in your field — the kind that allows you to evaluate AI outputs critically — is more valuable than ever.

What to Do With This Information

• Audit your current job for its most automatable tasks. Are you building expertise in those tasks or in the judgment-heavy tasks adjacent to them?
• Develop “AI oversight” as a skill — learn how your field’s AI tools work, where they fail, and how to catch their errors. This is the skill that pays a premium in virtually every industry.
• For career transitions: roles combining human judgment with AI tools (not replacing AI, not replaced by AI, but supervising it) are consistently among the most resilient.

Key Takeaway: AI is automating tasks, not jobs. The workers who suffer most are those whose jobs consist primarily of automatable tasks with little skill development beyond them. The workers who benefit most are those who develop genuine expertise and learn to work effectively with AI tools in their field.