One measure of our times is how difficult it is to give careers advice to our children.
A parent in the 1980s or 1990s had no difficulty. “Study hard, wear a shirt and become an accountant,” – good advice if your child wanted the stable life, some money and a pension. “There are always jobs in driving,” – again this was good advice back two or three decades. Parents had it easy. There might have been tantrums and fights about homework, but the path was always clear. Every kid knew that at the end of a few years of study and certificates, there were jobs in law, medicine, engineering, retail, cookery and so on.
This is not so now. One consequence of the development of Artificial Intelligence (AI) is that we do not know where the boundaries of machines lie and where humans take over. One thing for sure is that the machines will do more – much, much more – and that this will make some good jobs into a bad bet. The advance of AI will, in some profound way, affect every career that I have mentioned so far. That is accountancy, driving, law, medicine, engineering, retail and cookery.
That’s right there are even robot chefs.
We can see all of this in the economic statistics of today. It is called the “hollowing out” of the labour market. The term asserts that AI affects some white-collar jobs even more than it affects many blue-collar jobs. It “hollows out” from the centre, rather than just automating manual labour.
The greatest signal that something is afoot comes with autonomous vehicles. People are already getting ready for self-driving cars and trucks on our streets. Google rolls Waymo fully autonomous vehicles through Californian streets every single day and no one gets hurt. The crash stats are tell us that AI is a much safer driver than is a human being. This is remarkable because just back in 2004 Frank Levy and Richard Murnane, two leading economists, told us that driving a vehicle through a busy junction was a task so loaded with judgment that a machine could never manage it. It seemed to mark the boundary between a rule-driven AI and a human capacity for complexity and guesswork.
They were wrong. Levy and Murnane were not stupid – far from it, they are gifted scholars – it is just that the tech has come on very rapidly in the decade and a bit since 2004. This is because of ‘Machine Learning’. This is a form of AI through which machines learn for themselves about patterns of objects and behaviour that they encounter.
Machine Learning is a revolution on its own but then is aided and abetted in its progress by other forms of AI as well as other developments in computing, like the Cloud, faster processors and cheap sensors. It has some phenomenal effects. Think about this – when one machine learns about a particular task, they all learn about it. This is a “hive mind” effect and it means, for example, when one autonomous car learns how to navigate Hyde Park Corner, they all learn.
There are many really good examples of the progress of AI. A frequent example is how IBM’s Watson and other programs have been shown to do a better job of diagnosing illness than experienced doctors. Think about that “hive mind” again: when one machine recognises a rare form of cancer, they all can do it – spontaneous and cheap improvements in healthcare.
There’s another way in which AI surfaces in the press. This is with dire warnings of machines taking over everything and endangering human life itself. Some serious commentators including Elon Musk and Stephen Hawking have given voice to this. There is plenty to be concerned about but this is for another day. The kind of intelligence the commentators particularly fear is an Artificial General Intelligence (AGI) as opposed to the task-specific AI of today. So far, we are nowhere near an AGI so this is not on the horizon. Then again, we have learned to be cautious.
In the meantime we all have the problem of advising our kids on their careers. Maybe your child has a natural interest in TensorFlow, Python and R: that might be good news and make the job-market easier to navigate. No doubt we’ll still have doctors in some way too – but there might not be as much value in nursing. There will probably be a premium paid for the best human-to-human carers and therapists. There might also be a safe space for some artists where only a human being will do.
The rest of us, and especially our kids, can expect AI to take up camp in our jobs and to change the nature of work as we know it.
Written by Peter Kawalek