AI will shape the future of your workforce, but your workforce will also shape the future of AI. Pessimists claim that AI means the mass destruction of jobs, while optimists see workers freed from boring tasks and new types of roles being created. Both are missing the point by failing to answer the real questions says Tania Kuklina of KPMG.
How will AI reshape and transform the workforce and the nature of work, and how long until we see the full benefits of the technology? In the last 80 years, technological advances, coupled with profound economic and societal change have destroyed many old jobs.
For example, a common job in 1940 was ‘computer’ (essentially a human using a mechanical calculator). These lost jobs have been replaced with new ones - 60% of all occupations in 2018 did not exist in 1940.
The human touch
If we break down jobs into their component parts, they comprise certain explicit as well as implicit skills. AI is getting better at performing some skills, and also brings new skills of its own, but it requires careful prompting to provide what humans would consider implicit knowledge, such as context, ethics, communication style, and approaches to problem-solving. In other words, many implicit skills of human employees must be ‘programmed’ into the AI.
Imagine a proposal to replace hospital cleaners with sanitary robots. Research is presented to demonstrate an improvement in cleanliness and costs. However, a good hospital cleaner won’t slavishly follow cleaning procedures. They use tact and judgement to avoid intruding at delicate moments for patients and their relatives. Implicit human skills like empathy, understanding and good judgment should not be overlooked.
The impact on jobs
Will new automation and augmentation capabilities of AI reduce job numbers? AI is more likely to outperform humans in jobs with a narrow range of skills. For jobs demanding a broader range of component skills, the challenge will be finding new equilibriums between human and digital workers as AI capabilities continually evolve and advance.
History also teaches us that predicted savings from miracle technology are not always borne out by actual experience – the so-called Jevons Paradox. For example, when LED lighting made incandescent bulbs obsolete, it was predicted that energy usage for lighting would fall dramatically. Instead, businesses found more things to illuminate and the nighttime world grew brighter and more garish. We may choose to consume the efficiency savings generated by AI in the form of more satisfying jobs, and new sources of employment.
Another aspect to consider is the practical usefulness of what AI can deliver. For example, AI can recap a meeting for you, document minutes and actions and so on. All fine in theory, but we have survived until now without formal notes documenting the majority of our meetings. While Microsoft Copilot is an excellent companion, a human pilot remains essential for running meetings.
AI can comment on computer code, can assist in debugging, and can assist the average computer programmer, but it doesn’t yet have the human problem-solving skills required to fully replace them. The difference in performance between AI-generated Python code and a top-level human programmer’s machine code can be months to milliseconds.
Risk management
There are other good reasons to keep a human in the loop – apart from their notorious tendency to confabulate or “hallucinate”, AI agents are constantly being attacked and ‘jailbroken’ – bypassing their ethical controls. With early adopters suffering, at best embarrassment, and at worst serious legal problems, human review of AI outputs seems more essential than sensible.
If every AI agent has a human boss then every employee will need basic management skills to oversee a digital workforce of various AI tools and bots. They will need to appreciate the strengths and weaknesses of their digital colleagues, and this will need to evolve as AI improves. Digital subordinates will be expected to produce work of a certain standard and the humans overseeing them must remain engaged and alert to review and sign off.
The upskilling imperative
It is very likely that existing and future talent will require broader skillsets. The culture of the workplace will need to change. Technology is evolving faster than people can adopt. Organisations will need to foster a culture of accelerated technology adoption and continuous learning.
Interestingly, AI can be used to assist companies in dealing with current talent shortages by helping them optimise what they already have. KPMG’s recent Skills Data project with FTSE 100 organisations, showed that 79% of open roles could be filled by upskilling existing talent (54%) or redeployment (25%). Hiring and promoting from within and investing in existing staff vastly reduce hiring costs and increase employee engagement and loyalty.
AI can be used to draft job descriptions, identify skills required for different roles, uncover relevant skills lying dormant within the organisation, and make recommendations for bridging skills gaps.
KPMG has already helped a client deploy a career chatbot to advise employees on internal job opportunities and how they can bridge skills gaps through training, or gaining different experiences. It is obvious that AI will transform organisations and jobs far beyond what we can imagine at present, and far faster than previous technological advances.
I began by stating that AI will shape the future of your workforce, but your workforce will also shape the future of AI. People must be at the centre of your AI journey, both as critical enablers and as potential barriers to its adoption.
Organisations whose people understand the advantages, limitations, and risks of AI, and who co-design solutions will be most likely to successfully adopt GenAI while maintaining human judgement and expertise.
Get in touch
At KPMG we understand the pressure business leaders are under to get it right on tech and AI.
To find out more about how KPMG perspectives and fresh thinking can help your business please contact Tania Kuklina of our AI team. We’d be delighted to hear from you.
