Every major technological revolution has displaced workers. The printing press eliminated medieval scribes. The industrial revolution ended cottage industries. The internet reshaped entire sectors of media, retail, and finance. But the speed of the current AI and automation wave is different in one critical way: it is targeting cognitive work, not just manual labour.
The World Economic Forum estimates that 85 million jobs will be displaced globally by automation by 2025, a figure we have now exceeded, while 97 million new roles will emerge. The net number is positive. But that aggregate hides enormous disruption in specific sectors, and the jobs created are not always in the same places, requiring the same skills, or accessible to the same workers as the jobs being lost.
Here is an honest assessment of which roles are most exposed, and what is growing in their wake.
1. Radiologists and diagnostic specialists
AI systems now match or outperform radiologists in detecting dozens of conditions from medical images. Google's DeepMind can identify over 50 eye diseases from retinal scans with accuracy matching world experts. In mammography, AI catches more cancers while producing fewer false positives. Radiology will not disappear, but the number of radiologists required to read the same volume of images will fall substantially. The replacement roles are in AI oversight, complex case review, and patient communication, tasks the algorithms do not do well.
2. Long-haul truck drivers
Autonomous freight is moving faster on highways than in cities. Highway driving, repetitive, predictable, high-speed, is exactly the environment where self-driving systems perform best. Several US companies are now running fully driverless freight on specific highway routes. The complete transition will take longer than early predictions suggested, but the direction is clear. The roles replacing long-haul driving are in remote vehicle monitoring, last-mile logistics, and fleet management.
3. Bank tellers and loan officers
AI can assess loan applications, flag fraud, and handle customer queries more accurately and at far lower cost than human bank staff. Most of the world's major banks have already reduced teller numbers by 20–40% over the past decade. The trajectory continues. The growing roles in financial services are in financial advising, cybersecurity, and relationship management for complex clients.
4. Data entry and basic accounting clerks
Any job primarily involving the transfer of structured information from one place to another is vulnerable. AI handles data entry, invoice processing, and basic bookkeeping with near-perfect accuracy. The number of accounting clerks in the UK fell by 34% between 2015 and 2025. The remaining roles are in financial analysis, tax strategy, and auditing, tasks requiring judgment, not repetition.
5. Call centre workers
Large language models now handle customer service queries with sufficient competence to replace the majority of call centre interactions. Companies including Klarna and many major airlines have already significantly reduced call centre staffing after AI deployment. Human agents are shifting toward handling complex escalations, complaints, and emotionally sensitive interactions, the edges of customer service that automation handles poorly.
6. Travel agents
AI travel planners now generate personalised itineraries, monitor flight prices in real time, handle rebooking, and manage complex multi-leg journeys with minimal human input. The number of travel agent jobs in most developed economies has halved since 2015. The surviving segment of the market is luxury, specialist, and business travel, where human relationships and expertise still add clear value.
7. Assembly line workers in manufacturing
Robotic automation has been transforming manufacturing for decades, but AI-guided robots capable of handling variable tasks, assembling different products, adapting to different materials, are accelerating the displacement beyond what earlier projections anticipated. China, despite being the world's largest manufacturing nation, is simultaneously the world's largest buyer of industrial robots. The growing roles are in robot maintenance, quality assurance, and advanced manufacturing design.
8. Newspaper and print journalists
AI is now writing basic news stories, sports scores, earnings reports, weather summaries, faster and at lower cost than human journalists. Several major wire services have been using AI for routine content since 2019. The remaining journalism roles are in investigation, analysis, narrative writing, and the kind of source-driven reporting that requires human presence and trust. These are also the highest-value roles, the ones that actually drive audience engagement and subscription revenue.
9. Paralegal and legal research assistants
Legal AI can review contracts, search case law, identify relevant precedents, and flag compliance issues at a fraction of the cost and time of junior lawyers. Major law firms have already begun reducing paralegal hiring as a result. The legal roles growing are in client advising, courtroom advocacy, and complex transaction management, the human-facing, judgment-intensive work that AI augments rather than replaces.
10. Cashiers and retail checkout staff
Automated checkout has been displacing cashiers since the early 2010s. The latest generation of frictionless retail, where shoppers simply walk in, take items, and leave while AI tracks purchases and charges accounts automatically, is making the checkout process itself obsolete. Amazon Go stores have been running this model since 2018. As the technology cheapens, the model will spread. The retail roles growing are in customer experience, product curation, and complex sales.
"The question is not whether your current job will change. It will. The question is whether you are positioned to move toward the new roles, or away from them."
The jobs being created
The roles growing fastest globally include AI trainers and auditors (who ensure AI systems behave correctly), healthcare support workers (as aging populations require more hands-on care that robots cannot yet provide), renewable energy technicians, cybersecurity specialists, and data scientists. Soft-skill intensive roles in education, therapy, creative work, and complex sales are also showing resilience, because they depend on human connection and judgment that automation cannot credibly replicate.
The common thread in the growing roles is that they require humans to do things that are either physically novel, emotionally demanding, or require complex judgment under uncertainty. These are the edges that AI consistently handles least well, and where human workers are likely to remain valuable for longer than headlines suggest.