AI is raising the skills threshold for entry-level work

AI is raising the threshold young people face when they enter work, even though the underlying abilities employers value remain familiar. Communication, judgement, resilience and teamwork still matter, while AI literacy adds to that existing set of expectations. The difficulty is that some of those durable abilities develop through work itself, just as AI puts pressure on some junior tasks and roles through which young people gain workplace experience.

That tension emerged during an education committee meeting on AI and EdTech in education. Rose Luckin, professor emerita at University College London, put the change in broader terms: “It’s perfectly clear the nature of work is changing and we have to help young people be prepared for that, but as previously, that’s not just about helping them use the tech,” she said. For employers, educators and policymakers, readiness for an AI-shaped labour market consequently involves both technology skills and the established process through which people become effective workers.

AI is expanding the entry-level skills threshold

The immediate workforce challenge is a wider set of expectations. Young people still need the abilities that made someone employable before widespread AI adoption, and they increasingly benefit from knowing how to use and judge AI. That raises the preparation required for entry, especially because access to education, technology and workplace experience varies widely. AI literacy will also matter differently across jobs, while employers can continue developing job-specific technical knowledge after someone joins an organisation.

The existing layer of preparation consists of capabilities employers have long needed. Oral communication, confidence, resilience, creative thinking, organisation, planning and teamwork were all identified as valuable for young people entering employment. Josh Hillman, director of education at Nuffield Foundation, describes these as “employability skills” and as “transferable skills that are developed partially through the education system but then partially through those skills in work”. His description matters because it places part of employability development inside employment itself.

That division between education and work shapes what each can reasonably provide. Education can build communication, organisation and the capacity to learn, while employment gives young people another environment in which to develop judgement and apply those capabilities to real responsibilities. Employers can then teach job-specific technical skills, developing workers around the systems and requirements of their organisation. AI adds another layer to that division of responsibility.

That added layer includes AI literacy, broader digital and technical competence, and what Hillman calls “foundational knowledge and judgement in AI.” He argues that education needs to “evolve to some extent” as these requirements develop. Students consequently need enough knowledge to make informed decisions involving AI while retaining a foundation that remains useful as technologies and jobs change. The distinction matters because preparation for AI use is broader than training for one current tool or occupation.

Forecasting exact occupational requirements would make that preparation brittle because, as Hillman put it, “[Education] can’t possibly predict what the future needs.” Luckin reaches a similar conclusion through the ability to learn: “We can’t predict exactly what their jobs will look like, but what we can predict is that they’re going to be really good at learning.” Rapid technological change makes adaptability and foundational judgement useful across possible futures. Those capabilities give employers a base on which to add the requirements of a particular role.

The employer’s role continues after hiring because occupational skills change with the job. A person with strong transferable abilities can learn job-specific technology at work, while foundational AI knowledge can help that person understand when and how an AI system should be used. Employers still have a training role because the skills needed for a particular job “evolve to some extent”; education cannot finish that process before employment starts. The problem is that employment opportunities themselves are also changing.

The entry-level paradox: young people need experience as AI puts experience-building roles under pressure

The skills threshold becomes harder to meet when changes in junior work affect the route through which people acquire those skills. Administrative work commonly allocated to junior staff can increasingly be automated with AI, putting some tasks historically available to inexperienced workers under pressure. Those tasks have value beyond their immediate output because work is one of the environments where people acquire transferable skills and become ready for greater responsibility. Removing or changing a task can consequently affect the development route as well as the work itself.

Hiring uncertainty creates a second pressure even when a task remains. Hillman describes entry-level jobs as “exposed” because firms facing uncertain conditions can be reluctant to commit to hiring, and that hesitation “hits entry-level roles first.” AI can affect junior opportunities by changing tasks, while uncertainty around the changing environment can determine whether employers recruit at all. Together, those mechanisms make entry routes more fragile without implying that every junior occupation is changing in the same way.

The result is a mismatch between expectations and opportunity. Young people are being asked to arrive with a broader base of readiness while work remains an important place for developing it. Kester Brewin, associate director at the Institute for the Future of Work, captures the demand side directly: “The basket of skills that young people are being asked to bring to a first job has grown.” The larger basket moves more preparation to the period before the first job.

Brewin makes that burden explicit: “The demand is bigger for them, which puts a huge onus on young people there,” he said. For an employer, a higher expected baseline may make sense when applicants arrive able to use current tools. Across the labour market, however, that baseline raises a practical question about how a person without previous workplace experience acquires the capabilities needed to secure that experience. Prior educational exposure to AI can make a material difference here.

Some graduates may enter in a stronger position because they already use AI during their studies. Hillman said: “Graduates probably on average are having more experience of using AI as part of their education, so they are acquiring some of the AI literacy, which means when they’re entering those roles, they are probably better able to immediately apply some of those in the workplace in a sophisticated way.” Educational exposure can consequently become a labour-market advantage before an employer provides training. That advantage makes differences in preparation more visible between candidates.

A graduate who has learned to apply AI appropriately can potentially contribute with it immediately, while another young person may be expected to acquire similar knowledge independently. The difference becomes especially important when access to work experience, technology and AI-enabled education is already uneven. Those differences also shape how families and young people interpret the changing market. The effects of AI are consequently felt before an application reaches an employer.

Parents are beginning to alter the career advice they give their children because of AI, while young people themselves believe the technology will make employment harder to obtain. Will Akrigg, UK government affairs manager at The King’s Trust, says AI is “changing the picture” for youth employment and having a “massive impact on the job market generally.” Those perceptions matter because they influence decisions about preparation and career paths before young people reach recruitment. The recruitment process can then reinforce the same sense of change.

That concern sits alongside a growing number of people in the UK who are NEET, meaning not in education, employment or training, although AI is one factor within a broader labour-market context. Recruitment adds another difficulty because AI is heavily used in hiring processes, which has been described as “disheartening” for young people searching for work. AI consequently reaches an early-career applicant through the skills expected of them, the work available to junior employees and the process used to select them. Each point can affect access to the experience from which later skills develop.

The size of those effects depends heavily on the job and sector. Hillman argues that “We’re probably overestimating the changes that are happening in the short term and are underestimating the long-term change, but the direction of travel is quite clear – AI is changing work across a wide swathe of occupations and sectors unevenly.” That unevenness shifts the workforce question toward identifying which tasks are changing and where inexperienced workers can still learn through employment. Careers guidance and employer planning both depend on that more specific view of changing work.

Education and employers are poorly aligned as the target moves faster

Changing tasks create a coordination problem because schools and employers control different parts of the route to work readiness. Schools cannot provide every occupational skill for every possible job, while firms train new and existing workers but also expect entrants to have some prior knowledge. Workplace requirements also need to reach the education providers preparing those entrants. AI makes that handoff harder because the required knowledge and practices are changing quickly.

The existing skills system struggles with that pace. Patrick Milnes, head of policy for people and work at the British Chambers of Commerce, said: “The skills system isn’t aligned with the pace of change that we’re seeing at the moment, so there is an urgent need to develop a serious plan for long-term AI literacy – and not just in education, but in lifelong reform of how people can adapt as the technology changes.” From the British Chambers of Commerce’s employer perspective, AI compounds the coordination problem because it is moving faster than earlier technology reforms and practices for using it are less well understood. That also makes the language used by education and employers important.

Milnes describes a gap in that shared language: “Education providers and employers think about skills, not just AI skills but all skills, in quite a different capacity, and that can make joining up the two groups quite difficult in terms of preparing young people for the world of work,” he said. His proposed direction reflects the employer interests represented by the British Chambers of Commerce: “We’d like to see more engagement with employers around all of this.” Greater engagement can make requirements clearer, but it cannot make future occupations fully predictable. Young people therefore also need current information to make their own decisions.

Hillman calls for careers guidance grounded in current labour-market information: “Young people need to have accurate labour market information about what jobs are changing, what tasks are changing within those jobs; they need to see the emerging areas where those jobs are going to be – do they need to move, do they need to think about recruitment practices?” That approach focuses guidance on changing tasks, geographic mobility and recruitment practices because an occupational title alone cannot capture how work is evolving. Careers systems can also make viable entry routes easier to find as those routes change. That role becomes more important as schools face demands extending beyond conventional academic knowledge.

Brewin describes the conventional school model as teachers passing knowledge to students and exams and grades measuring “how much of that transfer has happened.” Yet “schools are being asked to do so much,” he argues, while expectations around technology and employment continue to expand. The growing skills threshold consequently raises a question about what schools are expected to deliver within that model. AI brings the purpose of the model itself into the discussion.

For Brewin, “AI and other technologies are fundamentally changing that, so schools have to reimagine what it is they’re for”. The workforce implication is significant because adding another technical topic to an already crowded curriculum does not settle who should develop employability, AI judgement or occupational competence. Schools, employers and careers systems each control a different part of that process, so alignment between them becomes part of workforce readiness. Unequal access to those different parts makes the coordination problem more consequential.

AI readiness can become an inequality multiplier

The coordination problem becomes more serious when young people start from very different positions. Hillman argues that AI is “definitely narrowing the number of opportunities for those young people, and it’s also changing the nature of those jobs, but I think it’s particularly going to be challenging for those young people who have lower-level qualifications, or struggled to get work experience and work readiness through their pathways.” Because experience helps build employability, weaker access to experience can compound weaker access to changing technical skills. Digital access creates another dividing line in the same process.

Socioeconomic circumstances, disability and geography all affect access to technology across the UK, while effective AI use depends on reliable basic infrastructure. As Milnes put it: “If you can’t even get reliable internet, you’re obviously not going to be able to use AI in the same transformative manner.” Connectivity is consequently part of the route to AI readiness rather than a separate infrastructure issue. Government investment is attempting to widen access and capability.

The UK’s Digital Inclusion Innovation Fund totals £11.7m invested across 80 digital skills programmes, while a separate £200m fund is intended to upskill businesses and support AI adoption and scaling. The two interventions address different parts of the skills environment. The Digital Inclusion Innovation Fund broadens digital inclusion efforts, while the £200m fund builds capacity among employers that will increasingly shape how workers encounter AI. Educational access adds another part of that environment.

The UK government is also working on free-to-use AI tutoring tools for schools across the country. The aim is to make educational resources more accessible, including to students who lack technology at home or whose families cannot afford private tutoring. Wider access to AI-supported educational resources could reduce one difference between students, while the conditions in which students use those resources still shape what they gain from them. That distinction connects access to the wider social setting of learning.

Brewin describes this as a “socio-technical” approach, meaning technological capabilities and their social context have to be considered together. He is concerned that AI trials in schools could focus heavily on what the systems can do while giving insufficient attention to the relationships through which education happens. “[An] extremely powerful set of technologies could be dropped in [to schools] without really thinking about the social implications on the fundamentally important ecosystem of relationships that is so precious in schools within learning,” Brewin said. His concern places relationships alongside infrastructure and tools as part of effective access.

Neil Selwyn, professor at Monash University, makes the boundary explicit for disadvantaged learners. “I’m worried there is a tendency to see AI as a quick fix for students who are struggling in school,” he said. A tutoring system may widen access to a particular educational resource, while disadvantages surrounding a learner can also involve needs that require sustained human and institutional support. Those needs determine the role technology can realistically play.

Selwyn therefore places people and capacity at the centre of any intervention: “There’s no quick fix for learners who are from social deprived multiply disadvantaged backgrounds; I think we do need to invest in people and resources and time,” he said. His limit on the role of AI is equally clear: “Tech can be part of that in the background, but I don’t think technology is ever going to be the solution to any kind of social disadvantage issue.” For policymakers and education leaders, digital inclusion consequently covers the conditions in which technology is used as well as its availability. Those conditions include the opportunities and human support through which judgement develops.

Luckin’s experience gives that risk a final dimension. “Most of my time as an academic I believed AI would be a great tool for breaking down barriers, for greater inclusion, and it’s not rolling out that way at the moment, and we really need to be careful about that.” The relevant divide is broader than access to an AI tool: it includes connectivity, educational support, opportunities to practise, workplace experience and people available to help young people turn access into useful judgement. As the entry threshold rises, workforce planning needs to account for those conditions because they determine who has a credible route to reach it.

Key takeaways for leaders

• Raise entry-level skills preparation: AI literacy is joining communication, judgement, resilience and teamwork as a baseline for work. Employers can focus hiring on durable capabilities and build role-specific technical skills through training.
• Protect routes to workplace experience: AI is automating some junior tasks at the same time employers expect new entrants to arrive with broader skills. Workforce teams can identify where automation removes development opportunities and create alternative paths for early-career employees to build judgement and experience.
• Align education with changing work: AI is changing tasks faster than education and skills systems can track them. Employers can give education providers clearer signals about emerging requirements while investing in training and accurate labour-market information.
• Treat AI readiness as an access issue: Connectivity, technology access, educational support and work experience shape who can develop useful AI skills. Policymakers, educators and employers can pair AI investment with human support and practical opportunities to prevent existing disadvantages from widening.