Choosing a Career in the AI Era

The first article showed that the top 1% is an income band, not a job title. The six high-income-linked occupation groups we looked at all had more people working in them in 2025 than in 1996, but those headcounts do not tell us how many workers were in the top 1%.[1] [2] The second article asked who captures the value when AI changes the work: the company selling the tools, the business using them, or both.

For a young graduate, I would turn those ideas into a different question: what field would I be willing to understand deeply, and how could AI help me do better work in it? I would not start by searching for the job title AI cannot touch. No title comes with that guarantee.

What the Australian evidence says so far

Jobs and Skills Australia projects total employment to grow by about 2 million people, or 13.3%, between May 2025 and May 2035. It expects the biggest growth among Professionals and Managers, with care and community roles also adding jobs. These are projections based on current economic and labour-market trends. JSA explicitly says they do not include the future effects of generative AI.[3] [4]

The Australian evidence on young graduates is more reassuring than some overseas headlines. A 2026 Department of Employment and Workplace Relations review found that employment among 20–24-year-olds grew slightly faster than employment among people aged 25 and over after ChatGPT was introduced. The share of young graduates working in Skill Level 1 jobs—occupations that typically require a degree—rose from 51.1% in November 2022 to 52.2% in November 2025. The report did not find that occupational reshuffling had accelerated.[5]

That is a snapshot, not a promise about the next five years. The same review points out that young people can be affected early if employers slow hiring, because graduates are still building experience. The practical message is not “don’t worry about AI”. It is “don’t abandon a field based on a headline from another country”.[5]

JSA’s separate study of generative AI in Australia found that, so far, AI is more likely to assist people with their work than replace whole jobs. Routine tasks have more automation potential, but what actually happens also depends on whether organisations adopt the tools and redesign the work around them. A task can be exposed to AI without the whole occupation disappearing.[6]

JSA also describes how existing jobs can gain new tasks and specialisations, while some workers may move across roles or sectors as new combinations of skills emerge. That is why I would pay attention to how a field is changing, not only to the job title on a course prospectus.[9]

Fields where AI and human skill can reinforce each other

I would look for fields with real demand and work that still needs people to understand the situation, check the result, and take responsibility. The following are not “safe jobs” lists. They are areas where current Australian projections show demand and where there are sensible opportunities to combine human knowledge with AI tools.

Health and care. Health Care and Social Assistance is projected to grow by 23% by 2035. Within that broad area, JSA projects strong growth for physiotherapists, nursing support and personal care workers, and ambulance officers and paramedics. AI may help with paperwork, information summaries or planning. The worker still needs to understand the person in front of them, notice when something is wrong, and make or escalate decisions that affect someone’s care.[4]

Engineering, infrastructure and the physical world. JSA projects an increase of about 196,000 Technicians and Trades Workers by 2035, even though their share of the workforce may edge down as other groups grow faster. Construction and electricity, gas, water and waste services are each projected to grow by 12%. In these fields, AI could support design work, inspection, maintenance planning and analysis. The work still has to meet physical conditions, safety rules and what is happening on site. A useful path might combine engineering or trade knowledge with data, automation or project skills.[3] [4]

Digital, data and cybersecurity. AI and data tools need people who can build, connect, secure and check them. JSA projects growth in roles such as ICT Managers and Research and Development Managers. The World Economic Forum’s global employer survey also lists AI and big data, cybersecurity and technology literacy among the skills growing fastest. That does not mean every graduate should become an AI engineer. A professional who understands a real industry and can use digital tools responsibly may be just as valuable as someone who knows the tools but not the work they need to support.[4] [10]

Education, finance, law and other professional services. JSA expects Professionals and Managers to make up a larger share of the Australian workforce by 2035. Education and Training and Financial and Insurance Services are also projected to grow. AI can take on parts of research, drafting, summarising and routine analysis. The work will still need people to explain choices, understand clients, negotiate, teach, and carry responsibility for consequential decisions. In these fields, the advantage may come from combining a recognised professional foundation with the ability to use and check AI in day-to-day work.[3] [4] [6]

The growth figures are a guide to demand, not a promise of salary or an easy first job. They also do not mean a university degree is the only route. JSA expects more than 90% of projected employment growth to be in occupations at Skill Levels 1 to 4, which cover a range of post-school pathways. Its projections show that skilled trade and technical roles can add workers even when their share of the overall workforce is not rising.[4]

If you have already chosen a field

I would not start again just because AI has arrived. I would learn how the work is changing from the inside. Pick one recurring task in your field and ask three things: could AI do part of it, could AI help while a person checks the result, or does the work need a person to lead it? JSA’s research separates the potential for AI to assist a task from the potential for it to automate that task. It also stresses that these are measures of potential, not forecasts that a role will disappear.[7]

For example, an engineering student could test where AI helps with design options or maintenance information, while learning how to check the recommendations against safety and site conditions. A law graduate could use AI to speed up an initial search or draft, then build the skills to test the authorities, understand the client’s position and explain the advice. These are examples of questions to explore, not claims that every employer has already adopted those workflows.

The skills I would build are a mix. Learn the tools well enough to use them confidently, but also learn how they fail. Get better at checking facts, protecting confidential information, explaining decisions and understanding the people who rely on your work. JSA describes AI capability as important for everyone, alongside non-digital capabilities. The World Economic Forum’s global survey likewise puts analytical thinking, creative thinking, resilience and curiosity alongside technology skills.[8] [10]

If you are just starting out, seek placements, apprenticeships, projects or early roles where you can see how work is actually done. Graduates need more than a polished AI-generated answer. They need experience with exceptions, mistakes, customers, equipment and judgement. If routine tasks change, employers and education providers will need to make sure people still get those early learning opportunities.[5] [7]

A simple way to choose

Before committing to a field, I would ask myself three questions. Is the problem important enough that I want to keep learning about it? Does the work connect to a lasting need, such as health, infrastructure, security, education or business decisions? Can I picture AI helping with part of the work while people remain responsible for the result?

Then I would try the field before making a large commitment. Talk to someone doing the work. Look at current job advertisements. Try a short course, placement or project. Compare university, vocational and work-based routes. A career choice is stronger when it is based on what the work feels like and what employers need—not only on a forecast or the promise of a high salary.

The lesson I take from the first two articles is that the value is not in a job title or an AI tool by itself. It comes from knowing a field well enough to see where technology can improve the work, and being trusted to check what it produces. I would rather become a clinician, engineer, teacher, lawyer or business professional who can work confidently with AI than chase a supposedly AI-proof title that may not exist.

References

More writing · Previous: Article 2 · Start the series: Article 1 · About Gokul