Pressure & Time: How AI Could Reshape the Labour Market

A great deal is being written about AI, jobs, and the labour market. Our aim at Ascham Grindal is to provide a clear and grounded introduction to how we advise clients on this critical topic. The following is a simplified version that approach to advice.

Our stance is straightforward: AI and automation should be viewed as General-Purpose Technologies (GPTs). Despite the extraordinary examples of AI in action, and the equally large volume of low‑quality output, its economic implications are best understood in the same way we analyse previous waves of technological change.

This position sets us apart from those who claim “this time is different”, whether apocalyptic pessimists or the more breathless strands of tech evangelism. We hold this view because it enables us to avoid vague speculation and instead develop practical scenarios our clients can use. We are not hunting for the 1% edge case, but identifying the 50% scenario most likely to shape the real economy.

Pressure and Time: How Economies Absorb Technology

Economies evolve under the combined influence of pressure and time. Pressure comes from how firms deploy technology to change production, upgrade processes, and improve efficiency, ultimately driving productivity and profitability. Yet pressure alone does not determine outcomes. Structural changes also trigger responses from governments, regulators, institutions, and markets—and these take time to materialise.

In short: the real economy moves far more slowly than equity markets.

AI is moving quickly, and the technology is increasingly capable. But it cannot instantly permeate every layer of a complex economy. Perhaps in time it will be able to do everything; but if so, it raises deeper questions—such as who will buy all the things it produces? Our analysis therefore focuses on what we can reasonably advise on today. Below are three scenarios, all of which could emerge simultaneously across different sectors or time‑scales.

Scenario One: The End of the Race Between Education and Technology

Goldin and Katz’s The Race Between Education and Technology remains one of the clearest explanations of how skills, education, and wage differentials shape labour market outcomes. Their premise is simple: workers with higher levels of education earn more because technological progress has historically increased demand for educated labour. For much of the 20th century, this demand was met by expanding access to schooling and degrees. Since the 1980s, however, the supply of educated workers has not kept pace with demand. Wage gaps between graduates and non‑graduates have widened, and a growing share of the UK workforce hold qualifications below Level 3. At the same time, employers have increased their expectations for higher‑level skills.

AI could influence this dynamic in two ways:

1. Raising educational attainment.
   AI‑enabled learning tools could help improve outcomes across the population, supporting learners who may previously have struggled to access high‑quality instruction.

2. Offsetting the need for higher formal education.
   If AI augments workers’ capabilities in the workplace, it may lessen the requirement for degree‑level knowledge in some roles. Combined with improvements in schooling, this could broaden the pool of suitably skilled workers without increasing the number of graduates.

These effects would be complex, presenting challenges for the traditional career model. They may also reduce wages in certain sectors. But the point remains that we could see AI reducing the need for people to necessarily attain higher levels of qualification. This may create more specific demand for a quality level of education rather than a specific qualification level.

Scenario Two: Unwinding Polarisation and Rebuilding Middle‑Tier Roles

A related possibility is that AI could help recreate demand for occupations in the middle of the labour market. Historically, these ‘blue-collar’ roles that have been eroded over recent decades as the labour market has polarised between high and low skills/occupations. Previously, the very expansion of middle income roles was the route towards broad improvements in standards of living and prosperity and occurred when industrial manufacturing and service sectors expanded requiring more labour, when domestic and global factors led to deindustrialisation, and manufacturing focused on automation and high-end skills this link was broken.

I am not suggesting that AI will recreate those same blue-collar jobs, but it is possible for AI to shift demand in a similar way to scenario one. In essence firms would need to see the value of being able to use AI to augment the skills of the workforce, expanding demand for labour and rather than focusing on recruiting the highest skill, recruiting more broadly. For this to occur, several conditions would need to align:

1. Widespread adoption of AI by firms.
   Early displacement of jobs is likely—some predict that up to half of entry‑level roles could disappear. But as firms learn to use AI productively, new tasks and roles emerge, creating reinstatement and new employment opportunities. This is critical to creating more demand later.

2. Lower barriers to using AI tools.
   If workers with lower levels of formal education can effectively use AI systems, supported by improvements to public education and training, the candidate pool for mid‑tier jobs expands.

3. Sectors realising productivity gains sufficient to expand employment.
   Job creation requires profitable demand for labour. Historically, firms take time to integrate new technologies effectively. For the UK, the balance between AI development and AI diffusion will be crucial as firms that benefit from the spreading of AI across different sectors are likely to be able to create profitable edge. Early labour market effects will be concentrated in sectors using AI to reduce costs and defend market share. Broader gains depend on firms adopting AI to enhance production and organisational performance.

Scenario Three: A New Polarisation Between Trades and Everyone Else

A more pessimistic scenario is that AI transforms service-sector occupations so dramatically that much of the progress made by degree‑holders is reversed, simply because AI can perform their tasks more efficiently. Such a shock would have significant economic and social consequences, including reduced household purchasing power and potentially prolonged economic downturns.

However, a more plausible version of this scenario centres on the fact that AI can readily replace cognitive tasks, but not many practical or manual ones. Robotics faces significant constraints—particularly energy and battery life. Even the most advanced robots operate only briefly between charges. As a result, construction, engineering, and other manual trades remain far harder to automate.

In this scenario:

• Demand for practical, technical, and manual labour increases.

• These roles attract a growing wage premium due to their resistance to automation.

• Service‑based occupations—traditionally better paid—experience long‑term downward pressure on wages.

This is not a return to a simple “blue collar vs white collar” divide. Many trades require high levels of technical skill and substantial training. But while AI reshapes service roles, this scenario posits that the most robust job creation may emerge in the trades.

Post‑script: The Reality of a Tool

AI’s capabilities are often overstated. In practice, its utility depends on the balance between the time it saves and the quality of the output. Tools such as AI‑assisted illustration or design still rely heavily on the user’s ability to craft effective prompts and evaluate results. AI may eventually close this gap, but we are not there yet.

In essence, we still need to know how to use the tool—even if it can crack the nut for us.

AI presents two potential futures: one in which it truly is different this time, driving unprecedented substitution of labour and requiring wholesale shifts in economic policy; and another in which AI follows the familiar path of previous technologies, influencing growth and reshaping industries, but within manageable bounds.

On this we highly recommend: Jones, Charles. A.I. and Our Economic Future. No. W34779. National Bureau of Economic Research, 2026. https://doi.org/10.3386/w34779.