Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Artificial intelligence can now write reports, generate images, analyse information, create software code and handle routine customer enquiries. For some workers, it already feels like a useful assistant. For others, it represents a direct threat to their livelihood.
The difficult question is no longer whether AI will affect work. It is how deeply it will affect individual jobs—and how quickly that change will happen.
Think about your current occupation, not the entire economy. What do you honestly expect AI to do to your own job during the next five years?
Poll Question
Why this question is difficult
The effects of AI are not equal across every occupation.
A copywriter may worry that businesses can generate basic marketing text automatically. A software developer may use AI to produce code more quickly but still be needed to review architecture, security and logic. A teacher may use AI to prepare lessons while continuing to provide human explanation, encouragement and classroom management.
Even people with the same job title can experience AI differently. One company may use it to assist employees, while another may use it to reduce its workforce.
That is why the debate cannot be reduced to only two choices: “AI will take every job” or “AI will take no jobs.”
AI may replace tasks before replacing complete jobs
Most occupations contain several different tasks. Some are repetitive and predictable, while others require judgement, accountability, physical presence or human relationships.
AI may take over activities such as:
- Producing a first draft
- Summarising documents
- Organising information
- Answering routine questions
- Creating standard reports
- Generating basic images or code
- Scheduling and administrative work
However, automating several tasks does not necessarily eliminate the entire position. The employee may spend less time producing routine work and more time reviewing results, dealing with exceptions or communicating with people.
A 2026 study examining millions of interactions with Google’s AI tools found that people were using AI primarily for collaboration—including research, drafting and troubleshooting—rather than handing over complete occupations to automation. This suggests that assistance is currently more common than complete replacement.
Why workers are still concerned
The fact that AI usually replaces tasks first does not make job-security concerns imaginary.
If one employee can complete substantially more work with AI, an employer may eventually need fewer people. Entry-level positions may be particularly vulnerable when their traditional responsibilities involve research, first drafts, basic analysis or routine customer support.
A 2026 job-search survey reported that some respondents believed AI or automation had already contributed to their job loss. However, self-reported experiences cannot establish exactly how many positions were eliminated solely because of AI.
Companies may also describe ordinary restructuring or cost-cutting as “AI transformation,” making the true effect difficult to measure.
The disagreement is therefore partly about timing. AI may not be replacing entire occupations today, but workers are trying to decide what it will be capable of doing five years from now.
AI may create different kinds of work
Previous technological changes eliminated some tasks while creating new industries and responsibilities. AI may follow a similar pattern, although the transition could still be painful for people whose existing skills lose value.
Potential areas of growing work include:
- AI system supervision
- Output verification
- Data quality and governance
- Cybersecurity
- AI-assisted creative direction
- Model safety and evaluation
- Employee training
- Workflow integration
- Human-centred services
- Regulation and compliance
The important question is whether new opportunities will appear quickly enough—and whether affected workers will have realistic access to the necessary training.
Some jobs may be more resistant than others
Work may be harder to automate when it requires:
- Physical activity in unpredictable environments
- Responsibility for serious decisions
- Trust and long-term human relationships
- Empathy and emotional understanding
- Complex negotiation
- Leadership
- Original judgement
- Hands-on care
- Legal or professional accountability
Nevertheless, “harder to automate” does not mean “untouched by AI.” A nurse, electrician, teacher, lawyer or mechanic may remain essential while using AI for documentation, diagnosis support, research or planning.
What would make AI feel beneficial?
For many workers, attitudes may depend on who receives the benefit of increased productivity.
If AI helps an employee complete work faster, should that result in:
- Shorter working hours?
- Higher pay?
- More meaningful responsibilities?
- Greater company profit?
- Fewer employees?
- Lower prices for customers?
People may welcome AI as an assistant while opposing its use as an unexplained justification for layoffs or constant employee monitoring.
Transparency will therefore matter. Workers are more likely to trust workplace technology when they understand how it is being used and retain a meaningful way to challenge important automated decisions.
