Will AI make it harder for non-graduates to climb the jobs ladder?

{"text":[[{"start":null,"text":""}],[{"start":9.48,"text":"Welcome back to The AI Shift, our weekly exploration of AI’s impact on the world of work. This week, we’re looking at the progression routes people take through the labour market. This topic has received plenty of attention when it comes to young graduates. But what about people who didn’t go to university? Many of them enter the white-collar workforce too. So what does AI mean for their prospects to progress into better work?"}],[{"start":37.07,"text":"John writes"}],[{"start":38.72,"text":"There has been a flurry of recent research making the case that those without a degree may end up seeing just as much career disruption from AI as the graduate class. The latest example is a report published today by US non-profits the Brookings Institution and Opportunity@Work, which looks at occupational disruption through a new lens, viewing particular jobs not necessarily as ends unto themselves but as crucial stepping stones on a journey from low to high-paid work for non-college-educated workers."}],[{"start":71.27,"text":"Based on fine-grained analysis of data on wages, job moves and common skills shared between occupations, authors Justin Heck and Mark Muro show that while the skills and credentials graduates acquire at university generally allow them to skip the bottom rungs of the career ladder and land straight into respectably-paid junior professional roles, non-graduates rely on early-career job-to-job transitions to acquire the abilities and experience necessary to gain a foothold in the white-collar world."}],[{"start":104,"text":"Heck and Muro identify three groups of occupations that form the steps on this pathway: low-paid “origin” roles such as receptionists or data entry clerks, “gateway” occupations including administrative assistants and bookkeeping clerks, and finally better-paid “destination” jobs like accountants, paralegals and managers in healthcare and education."}],[{"start":128.35,"text":"Gateway roles are the key, providing young non-college workers with a route out of low-paid work and an opportunity to acquire transferable skills, while simultaneously creating a talent pool for white-collar employers. Customer service is a good example of one such occupation: not typically a career but a stepping stone; accessible to non-graduates while also facilitating the acquisition of skills that are valued in better-paid roles such as sales representatives and HR assistants. In all, the authors find that more than 23mn US non-college workers have transitioned through occupations on these pathways into higher-paid work over the past decade alone."}],[{"start":null,"text":""}],[{"start":173.75,"text":"The key implication in the context of AI’s impact on the labour market is that if one of these critical pathway occupations is disrupted, the effects will be felt not just in that occupation but also upstream and downstream. If customer service roles start being automated by AI, a critical stepping stone to white-collar work previously used by receptionists is taken away, harming economic mobility. At the same time, companies looking to hire their new intake of sales reps or HR assistants will find the pool of suitably experienced workers is dry."}],[{"start":211.84,"text":"Moving beyond theory, Heck and Muro show that based on real-world usage of AI for performing work tasks, occupations in their origin and gateway categories are especially exposed to disruption, with the transitions from receptionist to customer service representative and bookkeeping clerk to accountant already imperilled. In all, only half of pathways into higher-wage work for non-graduates are not highly exposed to AI."}],[{"start":241.28,"text":"The findings pair nicely with work earlier this year from researchers Sam Manning and Tomás Aguirre, which found that most of the well-paid knowledge work jobs thought to be highly exposed to AI are performed by people with diverse sets of valuable and transferable skills, who are thus well suited to adjusting and adapting to disruption. Manning and Aguirre found that when combining theoretical AI exposure and this measure of adaptive capacity, the most vulnerable jobs are precisely those identified in Heck and Muro’s non-college pathways: less specialised white-collar roles such as clerks and administrative assistants, where the threat of automation looms large and alternative career options are also limited."}],[{"start":289.58,"text":"Sarah, we’ve previously argued for using a more multi-faceted measure of occupational vulnerability that goes beyond theoretical task-based exposure. These analyses suggest that once one does that, the white-collar wipeout narrative looks a lot more shaky. Are you persuaded?"}],[{"start":306.93,"text":"Sarah writes"}],[{"start":310.01,"text":"This research gave me a sense of déjà vu, John. Because it’s not the first time people have talked about technology “hollowing out” roles in the middle of the labour market. Indeed, economists have long argued that the last wave of technological change (which was brought on by computers and the internet) predominantly automated away “middle rung” jobs in factories and offices. Here is MIT’s David Autor on the topic in 2024: "}],[{"start":null,"text":"

By making information and calculation cheap and abundant, computerization catalyzed an unprecedented concentration of decision-making power, and accompanying resources, among elite experts. Simultaneously, it automated away a broad middle-skill stratum of jobs in administrative support, clerical and blue-collar production occupations. Meanwhile, lacking better opportunities, 60 per cent of adults without a bachelor’s degree have been relegated to non-expert, low-paid service jobs.

"}],[{"start":340.21,"text":"This new Brookings research is a good reminder that not all of those jobs in the middle were — in fact — hollowed out. There might be fewer than there were in the pre-computer age, but they still exist, and they still provide stepping stones for people without degrees into better-paid work."}],[{"start":358.81,"text":"Although this research is focused on the US, the same is true in the UK, where about one-in-eight young people without degrees actually work in degree-level jobs — predominantly in those sorts of roles which you can still “work your way into” such as HR executives, sales managers and managers in the retail and warehousing sectors."}],[{"start":381.64,"text":"The question, then, is whether AI will finish the job begun by computers, and eliminate those final stepping-stone roles which help people without degrees to progress into better-paid white-collar roles. "}],[{"start":394.99,"text":"I think it’s a plausible and worrying possibility. But there is a counter-argument that AI could have the opposite effect. In this scenario, AI tools help workers without degrees to do more complex tasks which were previously the preserve of people with higher levels of education. Indeed, Autor advanced this case in his essay, arguing that AI could “assist with restoring the middle-skill, middle-class heart of the U.S. labor market” by enabling “a larger set of workers equipped with necessary foundational training to perform higher-stakes decision-making tasks currently arrogated to elite experts, such as doctors, lawyers, software engineers and college professors.”"}],[{"start":442.04,"text":"I don’t think the early evidence we’re seeing from the software profession supports this hypothesis. As we wrote last week, AI seems to be boosting demand for the most skilled and experienced professionals, rather than allowing people with less expertise to eat their lunch. But it’s still early days, and what is true for software won’t necessarily be true for every other role. What do you think, John, do you think it’s possible that AI will restore — rather than further erode — the middle rungs of the labour market?"}],[{"start":477.23,"text":"John replies . . . "}],[{"start":480.76,"text":"Thanks Sarah — so much of the AI jobs debate gets flattened into a single narrative, and it’s important to appreciate that there are solid economic arguments for the same technology having directionally different impacts. "}],[{"start":495.15,"text":"Earlier this week Harvard’s Seyed Mahdi Hosseini Maasoum and Guy Lichtinger outlined another helpful framework for thinking about this. Where AI enhances individual workers’ productivity it widens wage inequality by multiplying the returns to specialist skills and knowledge; where it reduces barriers to entry by eliminating the need for specialist skills it has the opposite effect. Interestingly, they find evidence that both dynamics are already playing out in measured AI usage, and that at the whole economy level they can to an extent cancel one another out."}],[{"start":532.48,"text":"Forecasting the long-term balance between the two is a fraught exercise, but my read of the evidence to date is that at least in the short term the former effect is likely to dominate. Thus far the heaviest users of AI are the most well-educated, who were already the most productive and well-paid, and benefit from autonomy in how they use the technology as well as more multi-faceted jobs that are hard to automate in one go. Heck and Muro’s non-graduate pathway jobs look more vulnerable on all counts."}],[{"start":568.4200000000001,"text":"Recommended reading"}],[{"start":null,"text":"

1. The Guardian ran an erudite and thought-provoking long read by Kevin T. Baker about AI, bureaucracy, and the US bombing of a school in Iran (Sarah)

2. US finance professor Gregor Schubert has some fascinating new research showing that AI is significantly increasing productivity for non-work household tasks (John)

"}],[{"start":null,"text":""}],[{"start":579.2600000000001,"text":""}]],"url":"https://audio.ftcn.net.cn/album/a_1775276303_5810.mp3"}