Teenagers are preparing for the jobs of 25 years ago – and schools are missing the AI revolution

The government has recently released its national youth strategy, which promises better career advice for young people in England. It’s sorely needed: for teenagers today, the future of work probably feels more like a moving target than a destination. Barely three years after ChatGPT went mainstream, the labour market has already shifted under young people’s feet.

In the US, job postings for roles requiring no degree have dropped by 18% since 2022, and roles requiring no prior experience by 20%. Administrative and professional service jobs – once key entry points for school-leavers – are down by as much as 40%^[1].

While headlines often warn of looming mass job losses due to GenAI, the reality is more complex. Jobs are not simply disappearing but transforming, and new kinds of jobs are appearing.

Research has^[2] projected that the adoption of new technologies will displace around two million jobs in the UK by 2035. However, this loss is expected to be offset by the creation of approximately 2.6 million new roles, particularly in higher-skilled occupations and healthcare roles.

Despite a transformed job market, OECD data^[3] from 80 countries shows that most young people still aim for traditional roles – as architects, vets and designers as well as doctors, teachers and lawyers – even as demand rises in digital, green and technical sectors. One-third of students in the OECD survey said school has not taught them anything useful for a job.

Students from more disadvantaged backgrounds^[4] are hit hardest. They engage less in career development activities, have less access to online career information and are less likely to recognise the value of education for future transitions.

Meanwhile, the very skills young people say they lack^[5] – digital skills and being informed, followed by drive, creativity and reflection – are the ones the labour market now demands.

The workforce challenge is, fundamentally, an education challenge^[6]. But schools aren’t keeping up with the world students are entering. Despite unprecedented labour-market change, teenagers’ career aspirations have not shifted in 25 years.

While older students and graduates often have networks or some workplace experience to fall back on, school-leavers do not. Yet they need to prepare for a future in which the labour market is changing faster than ever.

Future-proof skills

Young people are told they need “skills for the future”. But the evidence about which skills matter is messy, uneven and often contradictory.

A few things are clear, though. One is that digital and AI-related skills now carry significant premiums. Workers with AI or machine-learning skills earn more, and early evidence^[7] suggests that GenAI literacy can boost wages^[8] in non-technical roles by up to 36%.

Cognitive skill requirements have also surged^[9]. Critical thinking, prompt engineering – the ability to ask the right questions and provide clear, context-rich instructions to AI tools to obtain relevant results – and evaluating AI outputs are increasingly valued^[10].

However, not everything can be outsourced to AI – especially numbers. While large language models (LLMs) excel at text, they do not perform as well on quantitative tasks that involve pattern detection or numerical reasoning, although this may change with new LLM models^[12]. This makes strong numeracy a growing advantage for humans, not a declining one.

Creativity and empathy also matter – even though AI is everywhere. The future paradox is clear: young people are expected to adapt to AI systems while also offering the human qualities that machines cannot. They must be data-savvy and emotionally intelligent^[13], digitally fluent and genuinely collaborative.

It doesn’t help that even employers are confused. Many organisations, especially small and medium-sized businesses, may not fully understand which AI-related skills they need or how to identify them^[14]. This confusion shows up in job ads, which shape who applies and who is excluded.

My research with colleagues shows^[15], for example, that language describing jobs influences the gender and racial makeup of applicants. Ads emphasising flexibility and caring qualities tend to attract more women, reinforcing workforce segregation. If employers do not know what skills they need, or what signals they are sending, it is unreasonable to expect schools to fill the gap alone.

Identifying demand

The UK lacks a coordinated national labour market information system that could help schools, policymakers and employers see – in real time – where demand is emerging.

Preparing teenagers for the future cannot be left to a single careers lesson or a one-off talk from a visiting employer. Nor can it rely solely on career advisers operating in isolation.

A whole-school approach, supported by the wider employment and labour-market ecosystem, would make a significant difference. This means linking every subject to real-world skills and careers, and every student routinely encountering employers, workplaces and skills-building opportunities. Teenagers need up-to-date information^[16] and advice about higher education and careers, and support that challenges stereotypes and barriers.

This is not about telling students there is a “right” job or a single future path. It is about giving them tools to navigate uncertainty with confidence.

Young people need schools that understand the world they are entering, and employers who understand what they are asking for. Most of all, they need systems that recognise the future of work has changed – and help them change with it.

References

1. ^{^} as much as 40% (papers.ssrn.com)
2. ^{^} Research has (www.nfer.ac.uk)
3. ^{^} OECD data (www.oecd.org)
4. ^{^} disadvantaged backgrounds (www.oecd.org)
5. ^{^} young people say they lack (www.edge.co.uk)
6. ^{^} education challenge (www.ai-in-education.co.uk)
7. ^{^} early evidence (www.sciencedirect.com)
8. ^{^} boost wages (papers.ssrn.com)
9. ^{^} also surged (arxiv.org)
10. ^{^} increasingly valued (arxiv.org)
11. ^{^} MAYA LAB/Shutterstock (www.shutterstock.com)
12. ^{^} new LLM models (onlinelibrary.wiley.com)
13. ^{^} data-savvy and emotionally intelligent (www.nfer.ac.uk)
14. ^{^} how to identify them (www.gov.uk)
15. ^{^} research with colleagues shows (academic.oup.com)
16. ^{^} up-to-date information (www.gatsby.org.uk)