As artificial intelligence (AI) becomes better at performing narrow tasks traditionally done by humans, how will governments create policies to protect peoples’ livelihoods? Should it be up to governments to do so or will it create a new generation of entrepreneurs? Will AI create a new breed of welfare recipients or will it spur governments to dramatically reform welfare? Some have argued that, while some jobs will be displaced, new jobs will be created in their place.
Clearly, there are some incredibly beneficial applications of AI ‒such as saving lives by detecting disease earlier and more accurately than can a human doctor, but it is hard not to imagine that AI will accelerate income inequality by reducing the number of jobs available while creating additional educational, social, and financial obstacles to acquire the skills necessary to obtain the jobs of the future.
Imagine a truck driver who has been on the job for 10 years suddenly loses his job because the trucking company he works for replaces its employees with self-driving trucks. What policies will a government have in place to ensure that this truck driver can learn new skills for the new economy, assuming there are enough jobs to go around? Even if the government were to have some sort of program for displaced workers to return to school and learn new skills, will all of them actually do it? This does not make the truck driver a bad person or necessarily lazy, but with their identity and self-worth demolished, not all will react in the same way at the same time. This could lead to a number of related social issues, such as rising levels of alcoholism, drug abuse, domestic violence, obesity, or other health problems.
Ironically, truck drivers have been in demand in the United States and their salaries have been on the rise in recent years, with their wages having cumulatively increased by 10.8% from January 2014 to February 2018, due to higher demand from people purchasing goods online, but this will not last forever. The upward salary trajectory of truck drivers will likely create more of an incentive for companies to automate this function by adopting self-driving trucks (at least for longer routes, such as traveling on expressways between cities). Local deliveries in cities will most likely still need to be completed by human drivers, at least for now. Drones are already beginning to supplant individuals in the package delivery business for critical medical supplies, especially in countries with less regulatory oversight than in Western countries, such as Rwanda. So, while AI is not likely to completely replace truck drivers, it could reduce the number of drivers needed, and consequently push wages down due to increased supply and lower demand.
Manual workers are not the only ones at risk from artificial intelligence; knowledge-based workers can also be displaced. A 2018 study demonstrated that an AI platform was more skilled at identifying issues in legal contracts than twenty experienced lawyers. The participants were given four hours to review five non-disclosure agreements and identify thirty legal issues, which they were subsequently scored on. The human lawyers achieved an average accuracy rate of 85%, while the AI outperformed them by attaining 95% accuracy. While each lawyer spent an average of 92 minutes to complete the task, the AI took only 26 seconds. So, not only was the AI more accurate but, simply based on the time required to complete the task, it would take approximately 212 human lawyers working in unison to match the output of the AI.
While some leading Silicon Valley technology titans have suggested that universal basic income could be one way to address the mass displacement of workers, it misses an important point about human nature. For the truck driver (as for most workers), it is not just about money, but also about dignity, having something to do and being able to take care of your family.
As children, many of us may have observed one or both of our parents (or caregivers) going to work every day and instilling in us the belief that hard work was a good thing and that it built character. The puritan work ethic was thought to be one of the foundations of what was good for the individual and society in general. This notion still pervades our collective consciousness in the United States today, even outside a religious context. The working day occupies our time and minds and gives us a sense of purpose. When that is taken away, are we really ready to explore what life could be like doing something completely different from what we may have known our entire lives?
Doing so is not simply about income replacement. It is also about activity replacement or a completely new vision of how we want to live in society. The German philosopher Friedrich Nietzsche made the famous and controversial assertion that “God is dead.” Nietzsche was highlighting that the age of the enlightenment shattered our traditional conception of God through the rise of reason and science. Nietzsche saw the death of God as the destruction of the moral foundation on which Western civilization was built.
If reason killed God, then will AI kill work? When work is taken away from us, even some of us, what will we do with our time? Will we all become philosophers pondering the meaning of life or will we chase fleeting pleasures in the smart cities of the future? Unfortunately, we are in uncharted waters and it is not clear what the true ramifications of widespread unemployment, if it were to occur, or what it would mean for the future. While providing some sort of social safety net in the form of a basic income may be necessary, there are much deeper issues at stake.
We must recognize that a lot of free time and a lack of purpose can bring out the best and worst in human nature. We need to start thinking about what we value outside of work and productivity if it becomes less important in the future. It is the choices that governments, companies, and citizens make that will determine our destiny with AI, and if it awakens the angel or beast lying dormant within us all.
Editor's Note: A version of this article first appeared here in International Policy Digest on October 3, 2018.