In light of the significant impact artificial intelligence (AI) is set to have on the world of work, we frequently speak with our clients and associates to compare viewpoints and find out how they are preparing for the changes ahead.
We recently spoke with Michael Rendell, Partner and Leader of Global HR Consulting Practice at PwC and Dr Nicola Millard, Head of Customer Insight & Futures in the BT Global Services Innovation Team. We’ve shared their interviews below.
1. What is your view on AI/automation and the likely effect this will have on the future World of Work?
AI and automation has the potential to make every worker 50% more efficient – achieved through simplifying the repetitive types of tasks that at the moment humans are relied upon to complete. It started with physical robots in manufacturing and with the development of AI the opportunity now extends to human led processes. This will touch all industries and professions and, unlike previous technology revolutions, will impact on all elements of the value chain. The effect will be most profound on those so called “white collar” jobs that have seen less change in the past – the “privilege of expertise” is under threat and with the democratisation of knowledge, the focus shifts to how experts add value through judgement rather than as a simple consequence of technical knowledge.
This shift has the potential to make the world of work much more satisfying and rewarding – as “intelligent machines” augment human activity and lift the burden of the uninteresting and repetitive activity. Collaboration between artificially intelligent machines and humans is the key. The types of work will change – many jobs will disappear but new roles will emerge. We will also see the development of more shared roles – partly carried out by a person and partly by an intelligent machine.
Implications of AI
2. As the impact of AI on the workforce increases- what key factors will be critical to the success of both individuals and employers alike?
Adaptive resilience – accepting the changes, while maintaining fall back arrangements that ensure critical processes can continue even if the AI fails to deliver. For individuals that will mean a period of dual skills – maintain the skills needed for the old ways, while developing the skills needed for the future. There will be a clear differentiation by age – the young will feel compelled to adapt, while the older workers will offer the resilience. The key attributes for employees in the future will be an ability to collaborate (both with others and with intelligent technology), agility and a hunger for change.
Successful employers will be those who are comfortable with disruption, focus on the tools to ensure successful collaboration and those that blend the best technology with the very best human skills. The key will be a clear understanding of what in their value chain adds real value and which components of this value adding activity is best performed by their employees (judgement, relationships, creativity, innovation etc) and by intelligent machines (faultless reputation, large data set analysis and pattern spotting, etc). Get this combination right and enterprises will thrive. For employers the goal will be achieving the right staff mix, and continually adjusting that mix as AI solutions are able to deal with more business processes.
3. As we adapt to new ways of working – what responsibility do employers have to ease the transition for employees?
In the adaptive resilience model, employers will need to alter their resourcing models – the most critical part of that will be retaining experienced staff through reskilling and retention for the period of transition – which could be 10 years. All of this will be happening at a time of staff becoming available on an individual contracting basis – but that should present no problems to the business that has all their expertise tied up in AI and automation solutions – just like the manufacturing robot – the developed expertise can never leave the building.
Enterprises in segments where AI will eliminate many roles have an obligation to explore how they turn their financial capital into social capital and find new ways of adding value to all their stakeholders (including the communities they support). This represents the biggest opportunity and challenge of the next 15 years.
4. What mind set do leaders need to successfully navigate this period of uncertainty and change?
For many enterprises AI represents more than just another technology development, business change or business process re engineering project. AI today and future development, has the power to change profoundly traditional business models and ways of creating value. We have been through the “hype cycle” and the change is real, today and into the future. The technology is getting exponentially less expensive and exponentially more powerful. The critical issues for leaders to consider when navigating this period of change are:
- what data do I have and how do I access it (the rocket fuel for successful AI adoption is data)
- do I have the right skills in my organisation and if not where do I find them and how do I attract them
- how do I protect my IP
- how do I invest and what is the ROI
- who should I partner with and stay close to, and
- how do I make my organisation agile and able to adapt to this change?
The challenge is to think not about what is possible today but to envisage what might be possible in two years and head towards that future – “…skating towards where the puck will be rather than where it is..” is critical when navigating the changes driven by AI.
5. How is your organisation preparing for the coming changes?
We have catalogued what we do, how we add value to our clients and wider stakeholders. We have tested this catalogue against the impact (now and future) of AI and begun a series of “experiments” to change/adapt our business and resourcing models. Some of these experiments are quite profound – for example, we have our first artificially intelligent PwC consultant focused on one area of expertise, some are less significant, for example we are developing chatbots to deal with mundane day to day question (“how do I….) by staff. This work is helping us define a clear strategy – with a focus on adaption and flexibility rather than a deterministic future vision. We are also continuing to develop relationships with both the largest and the smallest (start ups) organisations working in this space to stay right at the turbulent edge of this exciting development.
1. What is your view on AI/automation and the likely effect this will have on the future World of Work?
I think that having more intelligent technologies that take away the so-called “dull, dirty and dangerous” are potentially an amazing boost to productivity. Having technology manage my inbox, drive my car and fill out forms in duplicate frees me to do the things that actually matter.
I think that we probably overestimate the capabilities of these technologies (at least in the short to medium term) to eliminate entire jobs. There are jobs that are certainly at risk (mostly in transportation) and there is precedent for job categories to be eliminated (remember the typing pool) but they are likely to be in minority.
The more likely consequence is tasks will start to be eliminated – things like rekeying data into multiple different back end systems, transactional customer service queries, simple press releases and repetitive manual work.
This leaves us humans to do the more complex, emotive and empathetic work – skills like negotiation, innovation, creativity, manual dexterity and caring become more valued than they maybe are today. These are all things that are hard to codify into a machine – and AI is only as good as the data that is available to it. The explosion in data (due to increase as we move into an era of the internet of things, smart cities & buildings and clouds of clouds), coupled with deeper learning capabilities means that AI has a lot to work with, but it doesn’t have all the puzzle pieces. The puzzle is completed by human intelligence.
In short, these technologies make us value more what it is to be human.
2. As the impact of AI on the workforce increases – what key factors will be critical to the success of both individuals and employers alike?
There are a number of factors that we need to consider about the future workforce.
One of the more challenging issues to consider is that the trend towards wage polarisation that has already started with globalisation. This trend has caused considerable political instability and is one of many factors provoking the rise of nationalism, as a disenfranchised electorate voice their dissatisfaction with the situation.
AI has the potential to push this further. It is true that individuals who have the skills that are valuable (which includes carers, negotiators and complex problem solvers) are likely to be reaping the rewards, whilst those without the skills or the ability to adapt may get trapped in zero hours contracts, low wages and long term unemployment. This causes a ‘hollowing out’ effect that is extremely undesirable. Many solutions have been proposed for this – including a ‘universal minimum wage’ whether you are in employment or not, or a ‘robot tax’ to make the prospect of an entirely automated workforce less attractive to companies.
Another thing to consider is whether AI can start to provide employers with support for scarce skills, especially in science, technology, engineering and medicine. Simply importing these skills from other countries is likely to become increasingly difficult.
Medicine is a focus for AI developers because, to a certain extent, it is data driven (increasingly so as we start to wear and even ingest health monitoring devices). Capturing large data sets about health, combined with the deep learning capabilities of AI, can result in insights that help both clinicians and patients take real actions. This doesn’t eliminate the need to have clinicians – because they do A LOT more than just diagnostics – but it can be used to ensure that a clinician’s time with patients is optimised (especially as we have an aging population). The big challenge here is persuading patients to share their data with the AI in the first place.
One big clue to the future of AI can be found in Garry Kasparov’s experiences with IBM’s Deep Blue. The chess master was beaten by the machine in 1997. At that point he could have simply have thrown in the towel and given up. However, from that defeat came ‘advanced chess’ – where human and machine in combination are pitted against other humans and machines. As a result, an average chess player can become a supreme champion. Machines play chess in different ways to humans and vice versa – but the combination can create better outcomes.
Simply put, smart people partnered with smart machines have the potential to superpower us, if the combination is right.
3. As we adapt to new ways of working – what responsibility do employers have to ease the transition for employees?
For those employees on the sharp end of the automation process – and whose jobs are largely eliminated – there are a number of responsibilities that employers need to consider.
For example, workers on a manufacturing production line may well have their manual jobs automated, but there is still a need for people to supervise the production process and fix the machines when they go wrong. These may well be skills that the existing workers can be trained to do – but it requires investment from the employer and a willingness to learn and adapt from the employee.
It’s not just manufacturing, though. Big law firms are pouring money into AI as a way of automating tasks traditionally done by junior lawyers. Again, like clinicians, this will leave lawyers to focus on complex, higher-value work. The dilemma is how junior lawyers learn their trade in the first place, if they no longer have the easy stuff to cut their teeth on. This challenges how they are trained and mentored by their employers.
For employees who have more choice about accepting or rejecting AI, it will only work if employees actually embrace it.
Our experiences in the first wave of AI deployment in the 1990s illustrated this.
We deployed a neural network based knowledge system to help our customer service agents diagnose complex international private network faults. The idea was to allow them to concentrate on the customer, rather than the diagnostics. One deployment heavily involved the customer service agents in terms of both their knowledge and also figuring out what diagnostics worked and what didn’t. Another simply pushed data at them (and was also often wrong, but the agents had limited recourse in telling it that it was wrong). The first system worked pretty well, the second was largely ignored – neither were ultimately implemented because the technology was unwieldy and expensive to maintain at the time (the technology has moved a long way since then) but the lesson was learned: you can’t just drop technology on employees and expect them to embrace it wholeheartedly.
There is some horrible psychology underlying why we adopt or reject technologies. However, it boils down to 3 ‘U’s: Useful, Usable, Used.
Useful – We need to believe that the technology will actually help us do what we want to do (i.e. functionality) otherwise we won’t even consider using it. This is a belief rather than a reality – in reality, technologies typically do what they say on the tin. Our previous experiences will shape these perceptions (which is why it’s important to get new technologies right first time), as well as our confidence in using it and the perception of risk/accountability associated with the task. IT departments need to effectively “sell” the usefulness of AI to employees rather than just implement it and not tell anyone what it does.
Usable – we expect things to be both easy and intuitive to use and put us in control, otherwise we are unlikely to put the effort in to use them (unless we believe the effort is worth it). In terms of design, much of this falls under the banner of ‘usability’. Putting effort into the experience design process and trialling and testing things with employees can reap huge dividends in terms of getting new technologies like AI adopted.
Used – A technology that is not regarded as useful or usable is unlikely to be embraced by that many of us. Most of us need to actually successfully use and experience the technology to achieve our goals, otherwise we are unlikely to use it again. Peer influence effects (or ‘nudges’) can come into play here. Adoption is more likely to take place if our close colleagues or peers are using it. Leaders and managers need to lead by example by adopting the technology themselves.
4. What mindset do leaders need to successfully navigate this period of uncertainty and change?
Change is most definitely the norm these days, so leaders should already be used to managing uncertainty. I think that AI, especially, is getting overhyped and creating short term anxieties about things that are likely to happen in a much longer timescale. We are currently at the very start of the development of “cognitive AI” (i.e. AI that learns). What exists at the moment is “pattern matching” AI – which is already embedded in our lives every time we use services like Google, Netflix or Amazon. “Artificial General Intelligence” – which starts to match the capabilities and agility of a human – is still a good number of years away.
Our research work on employee morale shows that leaders can do 3 critical things to ensure that employees do not start to disengage with their work in the face of uncertainty.
- Recognise the value that their people bring to their work. People want to do work which is worthwhile and meaningful. They want to add value to their organisation and feel valued by the organisation in return. Telling them that they will be replaced by a machine probably isn’t the right approach here. Emphasising what they bring to the party and involving them in the change is a better way to go.
- Give employees a clear view of the future and a strong sense of progress towards that future. A sense that tomorrow is going to be better than today is crucial. People are able to tolerate far greater adversity if they can see that a situation is going to improve. Conversely, if people have no hope for the future, then morale will fall. It’s important to also understand that a lack of information about the future creates information voids. These are bad news. As Cyril Northcote Parkinson, the author of Parkinson’s Law, put it: “The void created by the failure to communicate is soon filled with poison, drivel and misrepresentation.”
- Create better relationships between employees. This is increasingly more important in a future where we are increasingly working alongside machines but also doing far more complex and challenging work. How we collaborate with each other and also with machines becomes central to employee engagement and productivity in the workplace. Leadership becomes less about command and control and more about establishing connections amongst employees and creating purpose for employees to collaborate.
5. How is your organisation preparing for the coming changes?
As Alan Kay once said: “the best way to predict the future is to invent it” – and this is very much the approach that we take to innovation. That’s why BT is a massive investor in research and innovation (the 3rd biggest in the UK). We have an innovation eco-system that comprises our own labs in the UK, Abu Dhabi, USA and China as well as university partners, tech scouts and customers around the world.
Research looking at applications of AI, the future of work, morale and collaboration are all exploring the practical impact of technologies like AI, cloud, the internet of things, collaboration and mobility on the ways that we do business. Rather than simply duplicate the analogue world into the digital one, we are looking at how we can reinvent the future world of work into the “digital possible”. This won’t work unless we develop different strategies from a people and process perspective, as well as from a technological one.