We have no idea what the job market will look like in 2050. It is generally agreed that machine learning and robotics will change almost every line of work – from producing yoghurt to teaching yoga. However, there are conflicting views about the nature of the change and its imminence. Some believe that within a mere decade or two, billions of people will become economically redundant. Others maintain that even in the long run automation will keep generating new jobs and greater prosperity for all.
So are we on a verge of a terrifying upheaval, or are such forecasts yet another example of ill-founded Luddite hysteria? It is hard to say. Fears that automation will create massive unemployment go back to the nineteenth century, and so far they have never materialised. Since the beginning of the Industrial Revolution, for every job lost to a machine at least one new job was created, and the average standard of living has increased dramatically. Yet there are good reasons to think that this time it is different, and that machine learning will be a real game changer.
Humans have two types of abilities – physical and cognitive. In the past, machines competed with humans mainly in raw physical abilities, while humans retained an immense edge over machines in cognition. Hence as manual jobs in agriculture and industry were automated, new service jobs emerged that required the kind of cognitive skills only humans possessed: learning, analysing, communicating and above all understanding human emotions. However, AI is now beginning to outperform humans in more and more of these skills, including in the understanding of human emotions. We don’t know of any third field of activity – beyond the physical and the cognitive – where humans will always retain a secure edge.
It is crucial to realise that the AI revolution is not just about computers getting faster and smarter. It is fuelled by breakthroughs in the life sciences and the social sciences as well. The better we understand the biochemical mechanisms that underpin human emotions, desires and choices, the better computers can become in analysing human behaviour, predicting human decisions, and replacing human drivers, bankers and lawyers.
In the last few decades research in areas such as neuroscience and behavioural economics allowed scientists to hack humans, and in particular to gain a much better understanding of how humans make decisions. It turned out that our choices of everything from food to mates result not from some mysterious free will, but rather from billions of neurons calculating probabilities within a split second. Vaunted ‘human intuition’ is in reality ‘pattern recognition’. Good drivers, bankers and lawyers don’t have magical intuitions about traffic, investment or negotiation – rather, by recognising recurring patterns, they spot and try to avoid careless pedestrians, inept borrowers and dishonest crooks. It also turned out that the biochemical algorithms of the human brain are far from perfect. They rely on heuristics, shortcuts and outdated circuits adapted to the African savannah rather than to the urban jungle. No wonder that even good drivers, bankers and lawyers sometimes make stupid mistakes.
This means that AI can outperform humans even in tasks that supposedly demand ‘intuition’. If you think AI needs to compete against the human soul in terms of mystical hunches – that sounds impossible. But if AI really needs to compete against neural networks in calculating probabilities and recognising patterns – that sounds far less daunting.
In particular, AI can be better at jobs that demand intuitions about other people. Many lines of work – such as driving a vehicle in a street full of pedestrians, lending money to strangers, and negotiating a business deal – require the ability to correctly assess the emotions and desires of other people. Is that kid about to jump onto the road? Does the man in the suit intend to take my money and disappear? Will that lawyer act on his threats, or is he just bluffing? As long as it was thought that such emotions and desires were generated by an immaterial spirit, it seemed obvious that computers will never be able to replace human drivers, bankers and lawyers. For how can a computer understand the divinely created human spirit? Yet if these emotions and desires are in fact no more than biochemical algorithms, there is no reason why computers cannot decipher these algorithms – and do so far better than any Homo sapiens.
A driver predicting the intentions of a pedestrian, a banker assessing the credibility of a potential borrower, and a lawyer gauging the mood at the negotiation table don’t rely on witchcraft. Rather, unbeknownst to them, their brains are recognising biochemical patterns by analysing facial expressions, tones of voice, hand movements, and even body odours. An AI equipped with the right sensors could do all that far more accurately and reliably than a human.
Hence the threat of job losses does not result merely from the rise of infotech. It results from the confluence of infotech with biotech. The way from the fMRI scanner to the labour market is long and tortuous, but it can still be covered within a few decades. What brain scientists are learning today about the amygdala and the cerebellum might make it possible for computers to outperform human psychiatrists and bodyguards in 2050.
AI not only stands poised to hack humans and outperform them in what were hitherto uniquely human skills. It also enjoys uniquely non-human abilities, which make the difference between an AI and a human worker one of kind rather than merely of degree. Two particularly important non-human abilities that AI possesses are connectivity and updateability.
Since humans are individuals, it is difficult to connect them to one another and to make sure that they are all up to date. In contrast, computers aren’t individuals, and it is easy to integrate them into a single flexible network. Hence what we are facing is not the replacement of millions of individual human workers by millions of individual robots and computers. Rather, individual humans are likely to be replaced by an integrated network. When considering automation it is therefore wrong to compare the abilities of a single human driver to that of a single self-driving car, or of a single human doctor to that of a single AI doctor. Rather, we should compare the abilities of a collection of human individuals to the abilities of an integrated network.
For example, many drivers are unfamiliar with all the changing traffic regulations, and they often violate them. In addition, since every vehicle is an autonomous entity, when two vehicles approach the same junction at the same time, the drivers might miscommunicate their intentions and collide. Self-driving cars, in contrast, can all be connected to one another. When two such vehicles approach the same junction, they are not really two separate entities – they are part of a single algorithm. The chances that they might miscommunicate and collide are therefore far smaller. And if the Ministry of Transport decides to change some traffic regulation, all self-driving vehicles can be easily updated at exactly the same moment, and barring some bug in the program, they will all follow the new regulation to the letter.
Similarly, if the World Health Organization identifies a new disease, or if a laboratory produces a new medicine, it is almost impossible to update all the human doctors in the world about these developments. In contrast, even if you have 10 billion AI doctors in the world – each monitoring the health of a single human being – you can still update all of them within a split second, and they can all communicate to each other their feedback on the new disease or medicine. These potential advantages of connectivity and updateability are so huge that at least in some lines of work it might make sense to replace all humans with computers, even if individually some humans still do a better job than the machines.
You might object that by switching from individual humans to a computer network we will lose the advantages of individuality. For example, if one human doctor makes a wrong judgement, he does not kill all the patients in the world, and he does not block the development of all new medications. In contrast, if all doctors are really just a single system, and that system makes a mistake, the results might be catastrophic. In truth, however, an integrated computer system can maximise the advantages of connectivity without losing the benefits of individuality. You can run many alternative algorithms on the same network, so that a patient in a remote jungle village can access through her smartphone not just a single authoritative doctor, but actually a hundred different AI doctors, whose relative performance is constantly being compared. You don’t like what the IBM doctor told you? No problem. Even if you are stranded somewhere on the slopes of Kilimanjaro, you can easily contact the Baidu doctor for a second opinion.
The benefits for human society are likely to be immense. AI doctors could provide far better and cheaper healthcare for billions of people, particularly for those who currently receive no healthcare at all. Thanks to learning algorithms and biometric sensors, a poor villager in an underdeveloped country might come to enjoy far better healthcare via her smartphone than the richest person in the world gets today from the most advanced urban hospital.
Similarly, self-driving vehicles could provide people with much better transport services, and in particular reduce mortality from traffic accidents. Today close to 1.25 million people are killed annually in traffic accidents (twice the number killed by war, crime and terrorism combined). More than 90 per cent of these accidents are caused by very human errors: somebody drinking alcohol and driving, somebody texting a message while driving, somebody falling asleep at the wheel, somebody daydreaming instead of paying attention to the road. The US National Highway Traffic Safety Administration estimated in 2012 that 31 per cent of fatal crashes in the USA involved alcohol abuse, 30 per cent involved speeding, and 21 per cent involved distracted drivers. Self-driving vehicles will never do any of these things. Though they suffer from their own problems and limitations, and though some accidents are inevitable, replacing all human drivers by computers is expected to reduce deaths and injuries on the road by about 90 per cent.8 In other words, switching to autonomous vehicles is likely to save the lives of a million people every year.
Hence it would be madness to block automation in fields such as transport and healthcare just in order to protect human jobs. After all, what we ultimately ought to protect is humans – not jobs. Redundant drivers and doctors will just have to find something else to do.
Write a reflection on the topic and hand it in by the end of the lesson (approx. 250 words). Reflect on the question: “We can potentially lose a lot of jobs in the future. How can you protect yourself from not losing your job to a machine with advanced AI in the future?