Artificial intelligence: will Terminator take over the workplace?
For many people today, the idea of Artificial Intelligence (A.I.) immediately conjures up the picture of the Terminator. This dystopian image betrays a deep-seated fear – that of being dominated, replaced, or even eliminated. It is our ancestral fear of wolves, and also our fear of others who are different and powerful. It also reflects our realisation that machines work tirelessly and without reward for humans, and that one day robots might be able to free themselves from their bondage.
Nightmares such as these could distract our attention from some practical issues relevant to this situation: the widespread use of A.I. is expected to “terminate” our socio-economic model and prepare the way for a new model to appear. It remains to be seen who will be eliminated or replaced, falling victim to this other form of Terminator.
Will Artificial Intelligence “terminate” full-time employment?
Employment is one of the major areas where this rupture is already apparent. Of course, E. Brynjolfsson and A. McAfee rightly point out that humans will not follow horses into obsolescence. However, there is a continual automation of repetitive work and machines are clearly more skilful than humans in a great many specialised areas ranging from data analysis to management. In addition to vehicle drivers and warehouse workers, one day A.I. will also replace radiologists, analysts and a lot of managers. Given the speed of this phenomenon, it is hard to imagine that enough new career jobs will emerge fast enough to give people work. We should therefore abandon the dream of full employment in the digital age and instead face up to the reality, in the most industrialised countries, of a society where work for humans will be scarce.
The areas where humans are least likely to be replaced by artificial intelligence are those where creativity and interpersonal skills are key. Reliance on such employment criteria means that inequality is likely to increase in our economies. What then can we do to preserve social ties? Obviously, there will be some simple work tasks for which it will be more cost-effective to employ humans. That is the case at the moment for manual sorting workers in certain mail order warehouses. But the tasks they have to carry out are still dictated by computer. Some machines work for humans, some humans are already working for machines.
While people might rejoice to see the definite end of heavy physical work, the workplace still has a role to play at different levels of the Maslow Pyramid: it represents a place where people can meet face to face and gives them a sense of dignity as well as income. While the general trend towards using A.I. is delivering the expected economic development, should this manna simply be distributed to allow every citizen to remain a solvent consumer? It might be worth reflecting on finding ways of rewarding activities that are useful to society and for which no economic model exists: shall we one day be paid for looking after our own children, for running artistic events or for visiting old people living in the same street? For that to happen, we would have to put away the concept of work in its purely economic context and replace it with the concept of an occupation. Machines will then have their place in society as complements to human activity.
Should humans be replaced or complemented?
It would still be most harmful if this evolution happened at the cost of losing skills in professional or personal domains such as driving vehicles, social relations or making life choices. In the near future, automation will increasingly become part of our lives. Will artificial intelligence render humans more intelligent or more stupid? Some people, like Joi Ito (http://joi.ito.com) are putting their money on a human-machine interaction: by developing the concept of society in the loop, the director of the MIT Media Lab is trying to create a virtual circle of mutual learning and develop collaboration between two forms of intelligence. We still have to take care that this approach can be adopted in a context often governed by a utilitarian philosophy for which the only output that matters is one that can be expressed in figures.
In the end, these decisions will lie with the artificial intelligence owners. But the resources needed for its development will be owned by a very small number of private individuals, mainly North Americans. Some countries in Asia have decided to compensate for this fact by investing massive amounts of public money. We await Europe’s decisions in this domain.
Eric Salobir o.p.
Translated from the original text in French
The views expressed in europeinfos are those of the authors and do not necessarily represent the position of COMECE and the Jesuit European Office.