I know what you’re thinking… Why is someone from tax at a talent conference?
Well, never fear: there’s a good possibility that advances in artificial intelligence mean people will soon no longer be employed in tax. You’ll never have to have someone from tax onstage again.
In fact, there’s a website you can visit – willarobottakemyjob.com – to confirm this.
See: for tax preparers, the probability of automation is 99%.
And the website also has this friendly message for me: You’re doomed!
OK, I knew you’d ask, what does it say about talent leaders?
For HR managers, there’s just 1% probability of automation.
Totally safe! That’s a relief. I’m just glad the numbers weren’t the other way around!
So why is it that this website and so many reports about the future of work suggest that tax preparers can be automated but HR managers can’t?
To answer this question, let’s get to grips with artificial intelligence.
Of course, the promise of AI extends far beyond CV screening into virtually every domain, sector, occupation and job.
With seemingly ‘superhuman’ capability, machines can play games of strategy, drive cars, talk to you, invest money, and create art. Is it any wonder that so many jobs are at risk of automation?
According to many pundits, to reduce the impact on jobs we should abandon specialised tasks and instead concentrate on our quintessentially human skills – like empathy, creativity, problem-solving and communication.
The problem with this whole narrative around technology and the future of work, though, is that it is dominated by a thread that puts technology before talent, that it’s more important to be competent at a task than to comprehend a task.
But have you ever stopped to wonder why humans have become the most intelligent species on the planet?
We’re all familiar with Charles Darwin’s theory of natural selection, which he presented to the RGS over 150 years ago. If we took this as the only template then intelligence should be governed solely by the number of neurons that evolution can pack into a cubic centimetre of brain, and we would be living in the trees with the other great apes while Whale Sharks, Elephants and certain species of crow ruled the Earth.
But human intelligence, our general intelligence, is the result of more than genetic evolution. Humans are the only species to pass on both DNA through our genes and knowledge through our language. Our intelligence is an incredibly sophisticated amalgam of naturally inherited competence, our ability to process data and information, and an amazing, culturally evolved ability to comprehend, to turn data and information into knowledge and wisdom and use it to adapt the world.
AI has no such comprehension. Nor, despite the hype, is it perfect; it can never be; it makes mistakes sometimes and often in unexpected ways. The argument that these mistakes don’t matter if the technology is better, faster, cheaper or safer than people overall only holds if you’re prepared to accept the possibility of catastrophic mistakes.
It’s true that some AI exhibits superhuman performance but this is typically at the risk of it occasionally being super-idiot. There are plenty of examples where this is a problem right now – I’m dealing with AI that is astonishingly accurate at classifying company transactions for tax purposes but which sometimes makes basic mistakes profoundly obvious to tax professionals. Meanwhile, others are dealing with self-driving cars that struggle to identify everyday road users, or autopilots that fight against a phantom stall.
If these possibilities exist, how can we trust AI? It’s not sufficient to say that people should focus on human skills while machines do the work. This kind of division of labour doesn’t make sense if a person still needs the knowledge and skills to correct mistakes when things go wrong. It’s also true that unless people are given opportunities to continue practising their skills, to DO the work themselves, their competence will degrade. When called upon in a crisis, neither technology nor human may be up to the job.
The challenge, I mentioned, with the previous definition of AI is that it suggests replication and competition with people.
Neither the machine nor its designers may appreciate that the way jobs are performed today is the result of decades of imperfect cultural evolution as well as design. Tasks may have mutated from generation to generation of worker through word of mouth, ad hoc training, user manuals or simple trial and error. The tasks are not what they seem.
However, when we think about replicating these tasks, we can suffer what Daniel Kahneman, the Nobel Prize-winning psychologist and economist, called “What you see is all there is”. In other words, we tend to look at a process and see only those aspects that are obvious. We can miss the implicit and intangible, and thus technology may become a dangerously imprecise and brittle replica.
The technology over talent narrative forces us to conclude that AI must inevitably decimate jobs, and will force people to focus exclusively on a narrowing strip of human-only skills that computers can’t yet mimic. But the idea that combining one or more highly specialised tasks creates a perfect replica of human work is utter fantasy.
There’s a different narrative, one that leads to a counter-intuitive conclusion:
Technology doesn’t eradicate knowledge and skills, it both broadens and deepens the requirement for them.
Think about it for a second: technology hasn’t removed the need for farmers to be good at farming; compared with ages past, when technology involved horses and a metal plough, today’s science and technology requires farmers to understand bioengineering and chemistry, geography, geology, meteorology and many other skills.
Greater use of technology has deepened the requirements for mastering farming, and the same is true of any occupation in which technology is embedded.
My conclusion is this: to survive in the future of work, people need more than just ‘quintessential human skills’. They still need to be T-shaped, but a T that gets broader and deeper with every new advance in technology.
These ‘Big-T’ people are specialised and multi-skilled. They have a good grasp of technology, appreciating what it does, how it works and where it may be strong and accurate, or biased and prone to error.
You are all Big-T people. You need to have knowledge and skills in psychology, neuroscience, organisational design, education, behavioural science and in the AI your organisation uses.
With ‘Big-T people’, organisations have a better chance of achieving the right balance between talent and technology.