Artificial intelligence has moved from being a futuristic idea to a practical workplace technology.

AI can now write reports, generate images, analyse documents, create software, answer customer questions, summarise meetings, translate languages, analyse data and perform increasingly complex digital tasks. AI agents are also beginning to move beyond simple chat and carry out sequences of tasks using software and other digital tools.

That naturally raises a difficult question Will AI take my job?

The short answer is: AI is likely to change a very large number of jobs, but that does not mean AI will simply eliminate a very large number of occupations.

For many workers, the first change will not be a robot taking their place. It will be software taking over some of the tasks they currently perform.

A writer may use AI to research and produce a first draft. A programmer may use AI to generate and test code. An accountant may use AI to analyse financial records. A lawyer may use AI to search documents. A customer-service employee may have AI handle routine enquiries.

The human worker may still be there but the nature of the job could be very different. That distinction matters.

Research from the World Economic Forum’s Future of Jobs Report 2025 suggests employers expect the boundary between human work and technology-assisted work to change substantially by 2030. Its employer survey found that 47% of work tasks were predominantly performed by humans alone at the time of the survey, while employers expected the balance to become much more evenly distributed between humans, technology and human-machine collaboration by 2030.

Meanwhile, UK government research published in January 2026 found that around 70% of UK workers are in occupations containing tasks that AI could potentially perform or enhance. Importantly, the report warns that exposure measures should not be interpreted as precise predictions of job losses.

AI Is More Likely to Change Jobs Before It Completely Replaces Them

One of the biggest mistakes people make when discussing artificial intelligence jobs is treating task automation and job elimination as the same thing.

They aren’t.

Imagine an accountant whose working day currently involves:

• Entering financial information
• Checking invoices
• Reconciling transactions
• Producing reports
• Analysing financial results
• Communicating with clients
• Advising management

AI might eventually perform much of the data-entry, reconciliation and reporting work.

That does not automatically mean there is no longer a need for an accountant.

Instead, the accountant could spend more time interpreting financial information, checking AI-generated analysis, advising clients and making higher-level decisions.

The same pattern can occur across many occupations.

Four ways AI can affect a job

1. Automation

AI performs a task previously performed by a person.

For example, an AI system might classify customer enquiries and automatically answer routine questions.

2. Augmentation

AI helps a human perform a task faster or better.

For example, a software developer might use AI to generate code but remain responsible for testing, architecture and final decisions.

3. Job redesign

The responsibilities of an occupation change because AI handles some activities.

A marketing professional might spend less time creating initial copy and more time developing strategy, analysing campaigns and managing brand positioning.

4. Job displacement

If technology can perform enough of the work cheaply and reliably enough, employers may need fewer people to perform that particular type of work.

This is the scenario people generally mean when they ask whether AI will take their job.

However, it is also the hardest outcome to predict.

How AI Actually Replaces Human Work

AI doesn’t need to be capable of doing an entire occupation to have a major effect on employment.

It only needs to perform enough economically valuable tasks to change how employers organise work.

Several technologies are contributing to this change.

Generative AI

Generative AI can produce:

• Text
• Images
• Audio
• Video
• Computer code
• Presentations
• Summaries
• Marketing material
• Structured information

This makes it particularly relevant to knowledge workers.

AI agents

AI agents represent a potentially more significant development.

Instead of simply answering a question, an AI agent can potentially:

1. Understand an objective
2. Break it into tasks
3. Use software
4. Search information
5. Produce outputs
6. Check results
7. Continue working through a process

As these systems become more capable and reliable, the amount of routine digital work that can be delegated to AI could increase.

Machine learning

Machine-learning systems can identify patterns in large datasets and make predictions or classifications.

This has applications in:

• Finance
• Healthcare
• Manufacturing
• Marketing
• Logistics
• Fraud detection
• Insurance
• Retail

Computer vision

Computer vision allows machines to analyse images and video.

It can be used for:

• Quality control
• Medical imaging
• Security
• Retail
• Manufacturing
• Agriculture
• Autonomous vehicles

Robotics

Robotics becomes particularly important when AI moves from the digital world into the physical world.

AI software can potentially control machines that perform physical tasks.

However, physical automation is generally more difficult than automating purely digital work because robots have to deal with unpredictable environments, safety issues, hardware costs and physical limitations.

The 50 Jobs Most Likely to Be Affected by AI

The following list should not be interpreted as a prediction that these 50 occupations will disappear.

Instead, these are occupations where significant portions of the work may be affected by AI, automation or AI-assisted productivity.

The distinction is important.

Some occupations may experience reduced demand.

Others may become significantly more productive.

Some may be redesigned.

And some may continue to exist while requiring completely different skills.

1. Customer Service Representatives

AI can already handle many routine customer enquiries.

Potentially automatable tasks include:

• Answering common questions
• Tracking orders
• Processing simple requests
• Categorising complaints
• Providing basic troubleshooting

Human workers are likely to remain important for complicated complaints, emotional situations, unusual cases and relationship management.

Potential impact: High task automation.

2. Data Entry Clerks

Data entry is one of the clearest examples of work that can be automated.

AI can extract information from:

• Documents
• Invoices
• Forms
• Emails
• PDFs
• Images

The role could increasingly shift toward checking, correcting and managing automated processes.

Potential impact: Very high task automation.

3. Administrative Assistants

AI can increasingly assist with:

• Scheduling
• Email drafting
• Document creation
• Meeting summaries
• Research
• Travel planning
• Data organisation

However, assistants who manage relationships, confidential information, complex logistics and executive priorities may continue to provide significant value.

Potential impact: High task disruption.

4. Bookkeepers

AI-powered accounting systems can automate transaction categorisation, reconciliation and routine reporting.

Human bookkeepers may increasingly focus on reviewing results, resolving exceptions and communicating with clients.

Potential impact: High task automation.

5. Payroll Clerks

Payroll involves many structured, repeatable processes.

AI and automation can assist with:

• Calculations
• Employee records
• Tax-related processes
• Payroll queries
• Error detection

Human oversight will remain important because payroll mistakes can have financial and legal consequences.

Potential impact: High task automation.

6. Telemarketers

AI voice systems could automate parts of outbound calling, lead qualification and appointment scheduling.

Human salespeople may remain more valuable for complex or high-value sales.

Potential impact: High task automation.

7. Translators

AI translation has improved dramatically.

Routine translation may become increasingly automated, particularly where:

• The language is well represented in training data
• The material is low risk
• Cultural nuance is limited

Human translators will remain important for legal, literary, diplomatic, specialist and highly sensitive content.

Potential impact: High task disruption.

8. Proofreaders

AI can identify:

• Grammar mistakes
• Spelling errors
• Repetition
• Inconsistent terminology
• Basic style problems

Professional editors may increasingly focus on meaning, tone, argument, originality and editorial judgement.

Potential impact: High task automation.

9. Copywriters

AI can generate advertising copy, product descriptions, emails and social media posts.

That does not necessarily eliminate the need for copywriters.

Instead, the value may shift toward:

• Strategy
• Brand voice
• Creative concepts
• Customer psychology
• Campaign development
• Editorial judgement

Potential impact: High task disruption.

10. Content Writers

AI can produce large quantities of informational content.

However, original reporting, expertise, first-hand experience, investigative work and genuinely differentiated analysis remain difficult to automate completely.

Writers may increasingly become editors, researchers, subject-matter experts and content strategists.

Potential impact: High task disruption.

11. Graphic Designers

Generative image systems can produce:

• Illustrations
• Concepts
• Advertisements
• Social media graphics
• Product mock-ups

Designers may increasingly compete on creative direction, branding, art direction and understanding clients rather than simply producing individual visual assets.

Potential impact: High task disruption.

12. Junior Software Developers

Software development is already one of the areas where AI usage is particularly concentrated.

AI can assist with:

• Code generation
• Debugging
• Documentation
• Testing
• Refactoring
• Code explanation

The impact on junior developers could be particularly important because some entry-level tasks are exactly the kind of structured digital work AI can perform.

However, software engineering itself is unlikely to disappear simply because AI writes more code. Architecture, requirements, security, system design and accountability remain important.

Potential impact: Very high task disruption.

13. Software Developers

Experienced developers are also affected.

AI can increase the amount of code an individual developer can produce.

This could change team structures and potentially reduce demand for certain types of routine development.

At the same time, more capable AI could increase demand for software by reducing development costs.

Potential impact: High task disruption, but continued demand for skilled developers.

14. Web Developers

Website generation tools can increasingly create:

• Page layouts
• HTML
• CSS
• JavaScript
• Content structures
• Basic functionality

Web developers may increasingly focus on complex applications, integrations, accessibility, performance, security and business requirements.

Potential impact: High task disruption.

15. Technical Writers

Technical documentation is highly compatible with generative AI.

AI can create:

• Documentation
• User guides
• API explanations
• FAQs
• Technical summaries

Humans will remain important for verifying technical accuracy and understanding complex products.

Potential impact: High task automation.

16. Market Research Analysts

AI can process large quantities of:

• Survey responses
• Reviews
• Customer comments
• Market data
• Competitor information

Analysts may spend less time collecting and organising information and more time interpreting it.

Potential impact: Significant task disruption.

17. Financial Analysts

AI can assist with:

• Data analysis
• Financial modelling
• Report generation
• Company comparisons
• Market research

However, investment decisions involve uncertainty, risk and accountability.

Potential impact: Significant task disruption.

18. Insurance Underwriters

AI can analyse large amounts of information and identify patterns associated with risk.

Underwriting may therefore become increasingly automated for straightforward cases.

Complex commercial risks may continue to require human judgement.

Potential impact: High task automation for routine cases.

19. Insurance Claims Handlers

AI can process documents, images and customer information to assess straightforward claims.

Humans may increasingly deal with disputed, complex or high-value cases.

Potential impact: High task disruption.

20. Paralegals

Legal professionals spend significant amounts of time reviewing and organising information.

AI can help with:

• Document review
• Legal research
• Contract comparison
• Summarisation
• Case preparation

Human legal judgement remains essential, particularly where consequences are significant.

Potential impact: High task disruption.

21. Legal Researchers

AI can search and summarise large volumes of legal information.

The role may increasingly involve verifying AI outputs and applying legal reasoning.

Potential impact: High task automation.

22. Recruitment Coordinators

AI can help:

• Screen applications
• Schedule interviews
• Match candidates
• Write job descriptions
• Communicate with applicants

Human judgement remains important when assessing culture, leadership potential and complex interpersonal characteristics.

Potential impact: Significant task disruption.

23. HR Administrators

Routine HR administration is highly digitised and therefore potentially automatable.

AI can assist with:

• Employee queries
• Documentation
• Scheduling
• Policy searches
• Reporting

Human HR professionals are likely to remain important for sensitive workplace matters.

Potential impact: High task automation.

24. Sales Development Representatives

AI can research prospects, create personalised messages and qualify leads.

Sales roles involving complex relationships may be more resilient.

Potential impact: Significant task disruption.

25. Digital Marketing Specialists

AI can assist with:

• Campaign creation
• Audience analysis
• Content generation
• Advertising optimisation
• Reporting

Marketing professionals may increasingly focus on strategy, positioning and customer understanding.

Potential impact: High task disruption.

26. SEO Specialists

AI can automate portions of:

• Keyword research
• Content briefs
• Competitor analysis
• Metadata creation
• Content generation

SEO professionals may increasingly focus on brand authority, technical strategy, original information and search behaviour.

Potential impact: Significant task disruption.

27. Social Media Managers

AI can generate posts, images, calendars and responses.

Human social media professionals will continue to provide value through:

• Brand strategy
• Community management
• Crisis communication
• Cultural understanding
• Creative campaigns

Potential impact: Significant task disruption.

28. Journalists

AI can generate straightforward reports from structured information.

However, journalism also involves:

• Interviews
• Investigation
• Source relationships
• Verification
• On-the-ground reporting
• Editorial judgement

The routine reporting layer may therefore be more exposed than investigative journalism.

Potential impact: Significant task disruption.

29. Financial Services Customer Advisers

AI can answer routine questions about accounts, products and transactions.

Human advisers may remain important for complex financial decisions and relationship-based services.

Potential impact: Moderate-to-high task disruption.

30. Customer Support Managers

AI can automate reporting, ticket categorisation and routine escalation.

Managers may spend more time handling exceptional cases, improving customer experience and managing AI-supported teams.

Potential impact: Moderate task disruption.

31. Purchasing Officers

AI can compare suppliers, analyse prices and monitor procurement information.

Human judgement remains important when negotiating relationships and handling strategic purchases.

Potential impact: Moderate-to-high task disruption.

32. Logistics Coordinators

AI can optimise:

• Routes
• Inventory
• Deliveries
• Scheduling
• Demand forecasting

Humans may increasingly manage exceptions and unexpected problems.

Potential impact: High task automation.

33. Dispatchers

AI can increasingly coordinate schedules and routes.

However, unusual situations can still require human intervention.

Potential impact: High task automation.

34. Receptionists

AI can handle:

• Appointment booking
• Basic enquiries
• Visitor information
• Phone routing

Reception roles involving customer experience and complex interactions may be more resistant.

Potential impact: High task automation for routine functions.

35. Travel Agents

AI can research destinations, compare options and build itineraries.

Human travel professionals may increasingly specialise in complex, luxury, corporate or highly personalised travel.

Potential impact: Significant task disruption.

36. Real Estate Administrators

AI can automate:

• Property descriptions
• Appointment scheduling
• Document processing
• Customer communications
• Market analysis

Human agents remain important for negotiations, relationships and transactions.

Potential impact: Significant task disruption.

37. Property Managers

AI can automate maintenance requests, tenant communication and administrative tasks.

Physical inspections, difficult tenant situations and property decisions may remain human-led.

Potential impact: Moderate task disruption.

38. Medical Transcriptionists

Speech recognition and AI can convert conversations into medical documentation.

Human involvement may increasingly focus on verification and correction.

Potential impact: Very high task automation.

39. Medical Coders

AI can analyse medical records and assign codes.

Because medical coding has regulatory and financial consequences, human review may remain important.

Potential impact: High task automation.

40. Radiology Support Roles

AI can assist with analysing medical images.

This does not mean radiologists simply disappear.

Instead, AI may act as another analytical tool while clinicians retain responsibility for diagnosis and treatment decisions.

Potential impact: Significant task augmentation.

41. Teachers

AI can create:

• Lesson plans
• Exercises
• Explanations
• Personalised practice
• Feedback

But teaching involves motivation, classroom management, relationships, safeguarding and understanding individual students.

Potential impact: Moderate task disruption.

42. Tutors

AI tutoring systems can provide personalised explanations and practice.

Human tutors may increasingly focus on motivation, mentoring and complex learning problems.

Potential impact: Significant task disruption.

43. Graphic Production Artists

Routine image manipulation and production work is increasingly automatable.

Creative direction and complex visual problem-solving remain more difficult.

Potential impact: High task automation.

44. Video Editors

AI can automate:

• Transcription
• Captions
• Cutting
• Scene detection
• Noise reduction
• Basic editing

Professional editors may focus increasingly on storytelling and creative decisions.

Potential impact: High task disruption.

45. Voice-Over Artists

AI-generated voices are increasingly capable of producing commercial audio.

Human voices may remain valuable where authenticity, personality, performance or licensing requirements matter.

Potential impact: Significant task disruption.

46. Research Assistants

AI can search, summarise and organise large amounts of information.

Human researchers remain essential for designing research, evaluating evidence and interpreting ambiguous findings.

Potential impact: High task disruption.

47. Junior Consultants

AI can perform research, analysis and presentation preparation.

This could change the traditional career ladder in consulting if junior employees spend less time performing routine analytical work.

Potential impact: High task disruption.

48. Tax Preparers

AI can automate calculations, document processing and routine tax preparation.

Complex tax planning and advice remain more difficult to automate.

Potential impact: High task automation for routine work.

49. Advertising Production Staff

AI can generate advertising concepts, images, copy and variations.

Human creative teams may shift toward campaign strategy and creative direction.

Potential impact: High task disruption.

50. Office Administrators

Many office administration tasks involve structured digital information.

AI can assist with:

• Emails
• Scheduling
• Documents
• Data
• Reports
• Internal enquiries

The role may evolve toward coordination and relationship management.

Potential impact: High task disruption.

The 50 Jobs at a Glance Effected By AI

Occupation	AI exposure	Likely effect
Customer service representative	High	Automation + augmentation
Data entry clerk	Very high	Automation
Administrative assistant	High	Automation
Bookkeeper	High	Automation
Payroll clerk	High	Automation
Telemarketer	High	Automation
Translator	High	Automation + review
Proofreader	High	Automation
Copywriter	High	Transformation
Content writer	High	Transformation
Graphic designer	High	Transformation
Junior software developer	Very high	Transformation
Software developer	High	Augmentation + transformation
Web developer	High	Transformation
Technical writer	High	Automation
Market research analyst	High	Augmentation
Financial analyst	High	Augmentation
Insurance underwriter	High	Automation of routine cases
Claims handler	High	Automation + escalation
Paralegal	High	Augmentation
Legal researcher	High	Automation
Recruitment coordinator	High	Automation
HR administrator	High	Automation
Sales development representative	High	Automation + augmentation
Digital marketing specialist	High	Transformation
SEO specialist	High	Transformation
Social media manager	Moderate-high	Transformation
Journalist	Moderate-high	Transformation
Financial adviser/support	Moderate-high	Augmentation
Customer support manager	Moderate	Augmentation
Purchasing officer	Moderate-high	Augmentation
Logistics coordinator	High	Automation + augmentation
Dispatcher	High	Automation
Receptionist	High	Automation
Travel agent	High	Transformation
Real estate administrator	High	Automation
Property manager	Moderate	Augmentation
Medical transcriptionist	Very high	Automation
Medical coder	High	Automation + review
Radiology support	Moderate-high	Augmentation
Teacher	Moderate	Augmentation
Tutor	High	Augmentation
Production artist	High	Automation
Video editor	High	Transformation
Voice-over artist	Moderate-high	Transformation
Research assistant	High	Transformation
Junior consultant	High	Transformation
Tax preparer	High	Automation
Advertising production	High	Transformation
Office administrator	High	Automation

Which Jobs Are More Resistant to AI

If you’re worried about artificial intelligence jobs and your own career, it is useful to understand what makes certain work difficult to automate.

Jobs tend to be harder to automate when they require combinations of:

Physical dexterity

Some jobs require manipulating unpredictable physical environments.

For example:

• Plumbers
• Electricians
• Carpenters
• Builders
• Mechanics

AI can help these workers with diagnosis, planning and documentation, but physically carrying out work in unpredictable environments remains difficult.

Human relationships

Jobs built around trust can be relatively resistant to complete automation.

Examples include:

• Therapists
• Nurses
• Social workers
• Teachers
• Coaches
• Senior advisers

Leadership

Managing people involves motivation, negotiation, conflict resolution and responsibility.

AI can provide information to managers but does not automatically eliminate the need for leadership.

Accountability

The more serious the consequences of an error, the more important human oversight can become.

This matters in:

• Healthcare
• Law
• Engineering
• Finance
• Safety
• Government

Ambiguous decision-making

AI performs particularly well when problems can be described clearly.

Real-world problems are often messy.

A manager might have to decide between two imperfect options with incomplete information.

That type of judgement can be difficult to automate reliably.

Jobs AI Could Create

The AI economy will not consist entirely of job losses.

New technology also creates new activities and occupations.

Some already exist, while others may become more common.

Potential areas include:

• AI engineers
• Machine-learning engineers
• AI implementation specialists
• AI governance specialists
• AI auditors
• AI security specialists
• Data scientists
• AI product managers
• AI trainers
• AI evaluation specialists
• AI safety professionals
• Robotics engineers
• Automation consultants
• AI integration specialists
• Human-AI workflow designers
• AI compliance specialists
• AI-focused cybersecurity professionals

The UK government’s 2026 AI Labour Market Survey found a significant skills gap in the country’s AI sector, with 97% of surveyed organisations reporting at least one AI-related skills gap. The survey also found increasing use of on-the-job training and growing demand for both technical and non-technical AI capabilities.

This illustrates an important point:

AI does not simply create a demand for people who can build AI models. It also creates demand for people who understand how to apply AI within real businesses.

When Will AI Start Replacing Jobs

Trying to provide an exact year when AI “takes jobs” is misleading.

There isn’t going to be one day when millions of jobs suddenly disappear.

The transition is more likely to happen in stages.

2026–2028: AI Becomes a Normal Workplace Tool

This stage is already underway.

Businesses are increasingly using AI for:

• Writing
• Research
• Customer service
• Coding
• Data analysis
• Marketing
• Administration
• Document processing

The most immediate effect is likely to be productivity and task automation.

Workers who use AI effectively may be able to accomplish more work.

At the same time, employers may reconsider how many people they need for certain routine activities.

2028–2030: AI Agents Become More Important

The next major shift could come from increasingly capable AI agents.

Instead of asking AI to perform one task, companies may give AI responsibility for sequences of related tasks.

For example:

Today:

Employee → AI → individual task

Potential future:

Manager → AI agent → research → analysis → software → report → review

This could create more substantial changes to office work.

The World Economic Forum’s 2025 employer survey anticipates significant changes in how work is divided between people and technology by 2030.

2030–2035: Workplace Structures Could Change

If AI becomes substantially more capable and reliable, companies may begin redesigning entire workflows rather than simply adding AI to existing jobs.

This could affect:

• Management structures
• Entry-level roles
• Administrative teams
• Professional services
• Customer service
• Software development
• Marketing
• Finance

However, this period is considerably more uncertain.

The capabilities of AI are only one factor.

Economic conditions, regulation, consumer preferences, organisational behaviour and the cost of implementation will all influence what actually happens.

2035 and Beyond: Highly Uncertain

Beyond the next decade, precise predictions become increasingly unreliable.

AI could become dramatically more capable.

Or progress could encounter technical, economic, regulatory or physical limitations.

Robotics could also become much more important if AI systems become capable of reliably controlling physical machines.

But nobody can responsibly tell you exactly which occupations will disappear in 2035 or 2040.

The useful approach is to prepare for multiple possible futures rather than rely on one prediction.

Is My Job at Risk With AI

If you are asking, “Is my job at risk with AI?”, don’t start by asking whether AI can perform your job.

Instead, examine your tasks.

Ask yourself:

1. How repetitive is my work?

The more repetitive the task, the easier it may be to automate.

2. Is my work primarily digital?

Digital work can often be automated more easily than physical work.

3. Can my work be described as a repeatable process?

If you can create a detailed set of instructions for completing a task, software may eventually be able to perform some or all of it.

4. How much human interaction does my work require?

Relationships, persuasion, trust and empathy can make automation more difficult.

5. Does my job require physical presence?

Physical-world work can be harder to automate than purely digital work.

6. What happens if AI makes a mistake?

If mistakes are inexpensive, automation may happen faster.

If mistakes could seriously harm someone or create major legal consequences, human oversight is more likely to remain important.

7. Can AI already perform parts of my job?

This is perhaps the most important question.

Don’t wait for AI to become capable of doing 100% of your job.

If AI can already perform 20% of your work, learn how to use it for that 20%.

Then determine what new value you can provide with the time saved.

A Simple AI Job-Risk Framework

You can think about your occupation using four broad categories.

Category	Description
Low current exposure	Most important tasks remain difficult for current AI
Moderate exposure	AI can automate or assist with several tasks
High exposure	AI can perform a substantial amount of digital work
Very high exposure	Much of the work consists of structured, repeatable digital tasks

This is not a prediction of whether you will lose your job.

Two people in the same occupation could face completely different outcomes.

An accountant who performs routine bookkeeping may experience much more automation than an accountant who provides complex strategic advice.

A software developer who writes repetitive code may be affected differently from an engineer responsible for designing critical infrastructure.

A teacher who spends hours producing worksheets may benefit enormously from AI without becoming redundant.

The occupation is only part of the story.

The tasks are what matter.

What Should I Do If AI Could Affect My Job

The worst response to AI is to ignore it.

The second-worst response is to panic.

A better response is to understand how AI can change the economics of your work.

1. Start Using AI

You don’t need to become an AI engineer.

Start by learning how AI can help with your existing responsibilities.

For example:

• Drafting
• Research
• Analysis
• Brainstorming
• Summarisation
• Data processing
• Meeting preparation
• Documentation
• Customer communication

The objective is not to use AI everywhere.

It is to understand where it provides genuine value.

2. Learn AI Literacy

You should understand:

• What AI can do
• What AI cannot do
• How AI makes mistakes
• How to evaluate AI output
• How to write useful instructions
• How to protect confidential information
• How to verify important information

AI literacy may become a basic workplace skill in the same way computer literacy became essential.

3. Become the Person Who Uses AI

One potential career advantage is becoming highly productive with AI within your particular industry.

A generic AI user is one thing.

A person who understands:

AI + accounting

or

AI + law

or

AI + construction

or

AI + marketing

can potentially provide much more value.

Domain expertise does not necessarily become less valuable when AI improves.

In some situations, it can become more valuable because someone needs to understand whether the AI’s output actually makes sense.

4. Strengthen Human Skills

AI makes some human skills more important, not less.

These include:

• Communication
• Leadership
• Negotiation
• Critical thinking
• Creativity
• Collaboration
• Empathy
• Decision-making
• Relationship building
• Problem-solving

The World Economic Forum’s Future of Jobs research similarly identifies a combination of technological and human capabilities as increasingly important through 2030.

5. Learn to Check AI

One of the biggest mistakes workers can make is assuming that AI output is automatically correct.

AI can produce convincing but incorrect information.

Workers who can:

• Detect errors
• Verify sources
• Challenge assumptions
• Identify risks
• Improve AI outputs

may become increasingly valuable.

6. Build Transferable Skills

Don’t build your entire career around one software product.

Instead, build capabilities that remain useful when technology changes.

For example:

Weak strategy:

“I know how to use AI Tool X.”

Stronger strategy:

“I know how to automate financial analysis using modern AI systems.”

The second capability is more transferable.

The Future May Be Humans + AI Rather Than Humans vs AI

The debate often presents two extreme possibilities.

Scenario one:

AI takes everyone’s jobs.

Scenario two:

AI simply makes everyone more productive.

Reality could be somewhere between the two.

AI can make individual workers dramatically more productive while also reducing the amount of labour required for certain tasks.

That can create a complicated economic effect.

A company might use AI to allow ten employees to do work previously requiring fifteen.

But lower production costs could also increase demand for the company’s products, potentially creating new work.

This is one reason forecasting total employment effects is difficult.

The World Economic Forum’s 2025 report provides an illustration: across the broader set of labour-market trends it studied, employers projected 170 million jobs created and 92 million displaced by 2030, for a net increase of 78 million jobs. Those figures are employer expectations across multiple structural trends, not a forecast of AI alone, so they should not be interpreted as “AI will create 78 million jobs.”

That distinction is important.

The Biggest Risk May Be the Changing Entry-Level Job

One of the less obvious effects of AI could involve how people start their careers.

Many professionals learn through relatively simple tasks.

A junior employee might:

• Prepare reports
• Research information
• Write basic code
• Analyse spreadsheets
• Answer routine customer questions
• Produce first drafts

Those are exactly the types of activities AI can increasingly assist with.

That creates a potential problem.

If AI performs many entry-level tasks, companies could need fewer junior workers to complete them.

UK data provides an early signal worth watching. A June 2026 government publication analysing LinkedIn data reported declines across many tracked entry-level occupations, including particularly large falls in accounting, graphic design and software engineering. However, the report also explicitly cautions that the observed patterns cannot yet be attributed directly to AI.

This is an important distinction.

AI may be contributing.

But labour markets are influenced by:

• Interest rates
• Economic growth
• Company finances
• Offshoring
• Industry cycles
• Education
• Demographic changes
• Regulation
• Technology

AI is only one variable.

What Will Happen to Salaries

This is another area where certainty is impossible.

AI could increase the productivity of highly skilled workers.

If a worker can accomplish twice as much using AI, their economic value could increase.

But if AI makes a particular skill much easier to obtain, the market value of that skill could fall.

For example, imagine that producing a basic marketing image takes a designer two hours today.

If AI eventually allows anyone to create a reasonable image in two minutes, the market for basic image production could change dramatically.

However, premium creative work could still command significant value because customers may pay for:

• Original ideas
• Brand understanding
• Taste
• Strategy
• Reputation
• Consistency
• Creative direction

Technology changes what is scarce.

And what is scarce often determines what the market values.

Three Skills That Could Become Extremely Valuable

If you’re concerned about your AI job prospects, focus on three broad capabilities.

Skill 1: AI Fluency

Know how to work with AI.

You should understand how to:

• Give instructions
• Break problems into tasks
• Evaluate outputs
• Use multiple AI tools
• Automate workflows
• Verify information

Skill 2: Domain Expertise

Know something valuable that AI doesn’t automatically understand in context.

For example:

• Accounting
• Engineering
• Healthcare
• Law
• Sales
• Construction
• Finance
• Education

Skill 3: Human Judgement

Know what should be done.

AI can generate options.

People still need to decide which options are appropriate.

The combination of these three capabilities may be more powerful than any one of them alone.

Frequently Asked Questions

Will AI take most people’s jobs?

There is currently no reliable evidence that allows anyone to state with certainty that AI will eliminate most jobs.

Evidence increasingly shows that AI can automate or assist with substantial portions of many occupations.

The size of the eventual employment effect remains uncertain.

What jobs will AI replace first?

Jobs containing large amounts of repetitive, structured and digital work are generally more exposed to automation.

Examples include data entry, routine administrative work, basic document processing and some forms of customer support.

However, exposure does not guarantee complete job elimination.

Is my job safe from AI?

No occupation can reasonably be described as completely protected from technological change.

A better question is how much of your work is exposed and how easily your responsibilities can change.

Which jobs are least likely to be replaced by AI?

Jobs involving complex human relationships, physical dexterity, leadership, accountability, unpredictable environments and high-stakes judgement may be more difficult to automate completely.

That does not mean they will be unaffected by AI.

Will AI replace office jobs?

Office work is among the areas likely to experience significant AI-driven change because much of it involves digital information.

However, office jobs are likely to be transformed at different rates.

Will AI replace programmers?

AI can already perform significant amounts of programming work.

That could reduce demand for some routine coding tasks.

But software engineering involves much more than writing code, including architecture, requirements, security, testing and system design.

Will AI replace teachers?

AI can automate parts of teaching, including lesson preparation, content generation and personalised practice.

Teaching also involves relationships, motivation, classroom management and safeguarding.

The occupation is therefore more likely to be transformed than simply eliminated.

Will AI replace accountants?

AI can automate many accounting tasks.

Accountants who focus on routine processing may face greater automation than professionals working in complex advisory, strategic or regulatory roles.

Will AI create more jobs than it destroys?

Nobody knows with certainty.

Some forecasts anticipate substantial job creation alongside displacement.

For example, the World Economic Forum’s 2025 employer survey projected a net increase in jobs by 2030 across the trends it examined.

However, that is not the same as proving that AI itself will create more jobs than it destroys.

What should I learn to protect my career from AI?

Start with AI literacy, then combine it with your existing professional expertise.

Develop:

• Critical thinking
• Communication
• Problem-solving
• Domain knowledge
• Leadership
• AI workflow skills
• Data literacy
• Adaptability

The goal is not to become impossible to replace.

The goal is to become highly valuable in a workplace where AI is widely available.

The Bottom Line: Will AI Take My Job?

AI will probably affect your job more than you currently realise.

But “AI will affect my job” and “AI will eliminate my job” are two very different statements.

Current evidence suggests that AI exposure is already widespread.

UK government research estimates that roughly 70% of UK workers are in occupations containing tasks that AI could potentially perform or enhance. At the same time, that research explicitly warns that exposure measures are not precise predictions of job displacement.

Real-world AI usage also shows a mixture of automation and augmentation. Anthropic’s Economic Index has found that AI usage has often involved humans working alongside AI rather than AI independently performing all of the work.

The most important change may therefore happen inside occupations rather than between occupations.

The accountant may still be an accountant.

The programmer may still be a programmer.

The teacher may still be a teacher.

The marketer may still be a marketer.

But what they spend their time doing could look very different.

The workers who benefit most may not necessarily be those who know the most about artificial intelligence.

They may be the people who understand how to combine AI with valuable human expertise.

So instead of asking only:

“Will AI take my job?”

ask yourself:

“What parts of my job can AI do?”

Then ask:

“What parts of my job are difficult for AI?”

And finally:

“How can I become better at the things that remain valuable?”

That shift in thinking turns AI from something you simply fear into something you can actively prepare for.

The future of work is unlikely to be as simple as humans versus machines.

It is much more likely to be a competition between different ways of combining people, technology, expertise and automation.

And that means the most important career decision may not be whether to compete with AI.

It may be learning how to work with it.