Which jobs will be replaced by AI – and which won’t

In this blog:

• Which jobs are truly at risk
• What AI can’t do – and what remains
• Why further training is crucial

There is a lot of talk about AI. Mostly in grand terms: disruption. Job killer. Silver bullet. As is so often the case, the truth lies somewhere in between. We are facing a profound transformation of the world of work. But not every change is a loss. And not every form of automation spells the end of a profession.

The reality is: AI is changing work. It is not replacing it across the board. Some tasks are disappearing, many are evolving. And some are even gaining in value.

Between fear and new beginnings

The debate is often characterised by extremes. On the one hand: the fear of mass unemployment. On the other: the promise of a more efficient, fairer world of work. In fact, the data paints a nuanced picture. Only 11% of jobseekers expect AI to have negative consequences for their careers. Almost 60% even see advantages. This confidence is remarkable – especially as the UN warns that up to 40% of all jobs worldwide could be affected.

So why are we so calm? Perhaps because there is significant investment in professional development here. Perhaps also because many companies view AI not as a replacement, but as a complement to human skills. That is a crucial difference.

Three types of AI – three types of change

What ‘AI’ can do is not uniform. It involves various technologies with different effects.

1. Generative AI (e.g. ChatGPT):

It generates text, analyses data and writes code. Particularly affected: clerical professions and administrative tasks. But highly skilled jobs – for example in the media or in software development – are also feeling the impact of this change. AI isn’t taking their jobs away. It’s changing them. Programmers are becoming system architects. Reporting is becoming faster, more analytical and data-driven.

2. Robotic Process Automation (RPA):

It takes over rule-based, repetitive tasks – for example in finance or HR. But here too, the jobs aren’t disappearing; they’re shifting. Those who used to enter data are now analysing processes. RPA takes the strain off – provided companies create new areas of responsibility.

3. Machine Learning (ML):

It recognises patterns, makes predictions, learns from experience. ML is fundamentally changing work, particularly in data analysis. Many entry-level jobs in this field are being replaced. At the same time, there is a growing need for specialists who can explain what algorithms actually do – and what they do not.

Who is really affected?

Entire occupational groups are not disappearing. It is tasks within jobs that are susceptible to automation. Particularly at risk are activities that…

• are repetitive
• follow clear rules
• require little contextual knowledge

Examples:

• Clerical professions such as bookkeeping or data entry
• Customer service, when limited to routine enquiries
• Transport & Logistics, due to autonomous driving and drones
• Entry-level jobs in data analysis
• Basic programming or dashboard creation
• Market research and sales at entry level
• Meat processing and other standardised manual tasks

And who is safe?

No one is safe. But a lot of tasks will remain human – at least for the time being. These include…

• Jobs involving high levels of social interaction (care, counselling, education)
• Tasks in unstructured environments (construction, trades, field work)
• Creative and strategic tasks requiring empathy or judgement
• Jobs involving ethical responsibility or legal assessment (judges, doctors, compliance officers)

The distinction, therefore, is not between ‘cognitive’ and ‘manual’. Rather, it is between ‘routine’ and ‘adaptive’. It is not the job title that matters, but the nature of the work.

Augmentation rather than replacement

A key point: many jobs will not be replaced, but transformed. This is called “augmentation”. AI takes over the routines – humans handle the exceptions, strategy and communication. This is more demanding. And it increases the pressure.

Those who use AI today often have more responsibility, more deadlines, and a greater flood of information. That is not always a relief. New skills are needed: in using tools, but also in managing stress.

Academic jobs are not standing still either

Academic professions used to be considered secure. Not so today. Whilst they are rarely fully automatable, many sub-tasks are. An example: lawyers still need to argue cases and pass judgement. But AI can produce initial drafts, carry out research and generate standard texts more quickly.

It’s similar in medicine. Diagnosis remains a human task – but what about image analysis or documentation? Here, AI is often more precise and faster. So anyone who thinks that a degree alone protects against automation is mistaken. It all comes down to the combination: specialist knowledge, digital literacy, ethical judgement.

The role of continuing professional development

It won’t work without learning. According to studies, 85% of employers want to invest in internal training. This is because almost 40% of the skills in demand today will be obsolete by 2030.

Two types of skills are in demand:

• Technical skills: data analysis, understanding of AI, programming, cybersecurity
• Soft skills: critical thinking, creativity, teamwork, adaptability

What is in demand are demonstrable skills – whether acquired through bootcamps, micro-certificates or on-the-job learning. Learning is becoming lifelong. And modular.

Who has access – and who doesn’t?

One problem remains: access to AI and further training is unevenly distributed.

• Women are more likely to work in clerical roles – and are therefore at greater risk. At the same time, they use AI significantly less often than men.
• Older employees make less use of AI tools than younger ones.
• Those with lower levels of education often have no access to suitable further training.

A new digital divide is looming here – not between humans and machines, but between those who use AI and those who do not (or cannot). Those who have access benefit. Those who do not fall behind.

Case studies 

Many  companies use AI – with varying results.

Industry:

The automotive sector is investing heavily in autonomous driving and process automation. Siemens, BMW, Bosch – they all rely on AI. Yet here too it is clear: it is not just about technology. When markets shrink, jobs are lost – despite AI. Productivity helps, but does not replace strategic direction.

Services:

In customer service, an IAB study has shown: AI-supported training makes new employees better more quickly. Satisfaction is rising. But two in ten felt more closely monitored. That, too, is part of the truth.

SMEs:

Here, the use of ChatGPT and other tools is growing. Yet access is often still limited. Many managers are optimistic – but only a third want to give all employees access to AI. That is not enough.

The future lies in the in-between

AI will not replace everything. But it is changing everything – step by step. The jobs that remain will look different. And they will demand more. More skill. More thinking. More humanity.

What we need now is a realistic perspective. No alarmism. But no romanticisation of technology either. It is about actively shaping AI – through education, regulation and bold companies.

Those who engage with AI early on, understand it and use it sensibly, will be better placed. Those who strengthen human skills – communication, ethics, creativity – will remain in demand.

And those who build systems that empower people, rather than replace them, create not only more productive but also fairer work.

Discover how Workday’s innovative solutions can help you navigate the evolving AI-powered world of work and future-proof your workforce.