The artificial elephant: Preparing for tomorrow’s workforce

11th December 2025
Getting to grips with the future of work has been a challenge for generations. Will advances in AI and automation change more than just the practice of work, and instead question even the concept of what we do and why? Read on to find out...

As shown by the commonplace practice of innovation and the adoption of creative and design thinking in business, we are quite capable of coming up with new ideas. Making decisions upon these generated ideas however can be trickier when trying to situate their value in an uncertain future.

While thinking about the future might seem complex, there are ways for us to think about the future using a simple framework. A framework where businesses can quickly think forward about the uncertainty of the next three-to-five years to help their decision making. Let’s walk through an example, the future of the workforce.

A big issue on people’s minds, as this is written, is the future of the workforce. There are many pressures on its size (numbers) and shape (skills and roles), and not just from technology, but also about the uncertainty of what else could happen in the world. We can use future thinking to de-stress that uncertainty and, by standing back from the minutiae of change, looking at a few simple possible futures and thinking about their implications and our options for actions under each.

Let’s talk about the challenge of sizing and shaping our workforce

A topic of some currency at the time of writing, the size and features of the workforce of the future in affluent countries, using a proxy sector of consumer facing organisationsi.

When we build out the future, we do it with breadth to include ideas about the economy, politics, social change and geopolitics rather than just technology.

Some uncertainties for business (with implications) includes

  • stagnating economies in the developed world (increasing costs and decreasing profits)
  • changes in consumer behaviour (decreasing demand)
  • demography (retiring population bulges, ageing population, limited skilled talent pool)

These and others are ALL pressures on the future of companies, the number of people employed and the mix of their skills.

The Artificial Elephant

The elephant in the room for many of these is Automation. People are talking about itii iii iv v vi vii but are unable to agree on the scale (numbers) and breadth (qualities) of its impact on the workforce; it is there but not there. We need to break it down to a manageable granularity.

Automation is a popular technology topic as it contains Artificial Intelligence, AI.  The spectrum of which includes

  • the pragmatic e.g. machine learning to automate repetitive tasks like medical scan interpretation.
  • enhanced or replaced human interaction e.g. natural language processing to help desk interactions. Conversational research.
  • generative for example code generation, language translation, media creation e.g. images and video.

Automation though is pointedly not just AI standing alone but also the effect of its discoveries. For example, drug discoveries and consequent life extension. There is also the swathe of technologies where AI will be an implicit enabler, such as the deep mathematics for fusion energy exploration, rather than as an explicit technology.

Like many other technologies of automation, AI can replace, augment and create work.  In terms of the type of worker, an MIT study on automation and labourviii used the differentiation of (a) tasks that are described using specialist terms  and (b) tasks described being made of simpler terms  We can play that out using a thought experiment of automating a piece of an organisation.

  • Automate simpler tasks and the number of workers doing them declines, the remaining tasks are more specialised and the associated pay often increases.

For example, if most of the simpler tasks at a London tube station are replaced by a combination of AI information, removal of journey charging, driverless trains and robotic cleaning. For the remaining workers doing specialised tasks, pay is likely to be increased.

  • Automate specialised work and the number of workers doing the remaining simple tasks increases but the pay is lower.

For example, if the specialised tasks at a London tube station  are replaced by onsite Agentic AI (being Agentic it is designed to develop local knowledge and operate without a dependency on network connectivity) the number of people doing the simpler tasks will increase but with lower pay.

The above expert / inexpert issue may be in the messy intermediate stage where AI is a technology that humans deploy. Geoffrey Hinton, known as the “Godfather of AI” made the assertion about this stage that:

“What’s actually going to happen is rich people are going to use AI to replace workers,” Hinton told the FT. “It’s going to create massive unemployment and a huge rise in profits. It will make a few people much richer and most people poorer.”ix

Beyond this lies the time that rather than humans, AI will deploy itself as it sees best fits.

Navigating this uncertainty

How can we navigate this impending transition? The complexity and uncertainty of the challenge of the future workforce means it is difficult to achieve resilience and adaptability using a forecasting approach with only one idea of the future. Using a future framework with multiple futures allows us more flexibility to look at future preparedness, resilience and options for multiple outcomes of automation. Futures, outcomes and strategies that vary widely from each other.

Let’s look at the possible futures

There are three advantages to using multiple futures here. First, the ability to involve the workforce themselves in their generation. Second, by seating the time of the target discussion in the future we step away from the tensions in the present. Third, by playing out concerns through using vastly divergent multiple futures, all can experience different viewpoints and step into the shoes of others.

It is worth pausing to think of the use of AI scenario production.  As we live in a world where human agency drives decisions, we look to balance the use of AI with the values of the people who will be affected by an issue and enacted change. In short, AI is useful for conversational research and, if required, generating images and video if desired. While it can generate scenarios, it is the social generation of them with unique mix of humanity, the stakeholders, leadership, management and workers, who need to participate in their creation that can be more useful in this workforce issue.

Our Futures

There are many possible futures for an organisation that range from ‘everything stays the same’ out to ‘acquisition’ and ‘discontinuation’. For our purposes, a set of futures for our workforce question focusses on the adoption of automation could be No Change, Managed Change, and Complete Changex  all set 10 years in the future.

Scenario Approach Some strategies
No Change to workforce required Pressures are worked through without any internal adjustment Workforce numbers maintained or increased to stay up with competitors who have evolved. Skill levels generally stable.
Some adoption of new technology with workforce upskilling and augmentation.
Managed Change to workforce Managed adoption of technology developments Workforce numbers reduced through natural attrition, e.g. retirement.
Adoption of new technology to compensate for attrition, to augment workforce and there is upskilling of remaining workforce.
Almost a Complete Change of the workforce Assertive adoption of new technology developments Workforce number reduction through negotiated strategies between workforce representatives and leadership. Investment in upskilling. Recruitment of new skills. Some use of outsourcing and right shoring. Some new hiring for new roles. Acquisitions. New business building

These simple futures are different to each and allow extremes of concerns to be played out. The futures are created, coloured in and played through collectively. This is important, especially for people who might be diametrically opposed to a future, as we can step more deeply into an oppositional one and help build it out for the best outcomes for all involved if it still occurred.

For example, someone who advocates No Change based on the current world can consider the Complete Change scenario. In doing so, they are invited to find opportunities for creative approaches. In exploring these scenarios as a group, stories can be shared about other industry transitions. A negative reference might be the closing of the coal mines and shipyards in the UK in the 1970’s and the problem transitioning an industrial workforce.  A positive reference for industry transition was in the 1950’s and 1960’s in the USA with the advent of containerised shipping. The dockworker unions in the West Coast of the USA negotiated a managed downsizing of the workforce with the foreseen arrival of containerised shipping from its success on the East coast.xii

It is important to note that, as people are involved, there will no doubt be mixed reactions to any major change. This will likely involve some confrontation but as this is a rehearsal of the future it informs us where work is required. In the case of the technological evolution of AI, a major change is inevitable for a commercial organisation as it must stay competitive with its industry peers.

How do we apply the findings?

Having generated the scenarios, there are two levels of strategy setting

  • High level to create a strategic approach to change e.g. retrain first, natural attrition strategies, looking for downsizing of units, creating radical change exploration projects – building an ideal picture of the business of the future
  • Low level to suit specific technology changes e.g. retraining schemes, early exploration programmes on early-stage technology

So what does this mean in practice?

The conclusion here is that strategy setting for major change does not have to wait for technologies or externalities to be settled. We can rehearse the future now and grasp the issue of reshaping the workforce by designing it for a business of the future rather than reacting to the winds of the present.

Who’s working on the problem at Imperial?

Professor Mark Kennedy: leading authority on how AI and automation are reshaping jobs and organisational structures. His research that models the intersection of technology adoption, automation, and job roles, providing data-driven insights into the future of work. His work emphasises that AI is an empowering tool rather than a replacement, highlighting the growing importance of human skills like creativity, leadership, and empathy in a tech-driven landscape.

References

back to all articles