The book The Economic Singularity proposes that Artificial Intelligence (AI), combined with related technological developments like the Internet of Things, will create mass unemployment as early as 2041. The author Calum Chace has penned several fictional and non-fiction books about AI and boasts 30 years’ business experience. Chace states that “a very large minority of citizens will be permanently unemployed”.
Everyone will be affected. One of the arguments for this conclusion is a study by McKinsey: “While automation will eliminate very few occupations entirely in the next decade, it will affect portions of almost all jobs to a greater or lesser degree, depending on the type of work they entail”. The research is ongoing, but some preliminary results have been released. One of the findings is that in the United States “as many as 45 percent of the activities individuals are paid to perform can be automated by adapting currently demonstrated technologies”. Many workplaces have, for example, administrative duties such as time tracking and filing business expenses that can be conducted faster and more accurately by machines. Imagine wearables (in or on the body) that automate time tracking. The wearables will know when you are in the office and at what time you leave. Or, you tell your phone or wearable that you are embarking on a business trip, so that all credit card expenses during that time will automatically be forwarded to your employer.
Handling emails could also, to some extent, be automated. All documents, emails, phone calls and tasks are saved in the cloud. Your AI assistant processes them and sends automated status updates to stakeholders. It also responds to pricing inquiries about your business services. When administrative duties can be automated, all jobs will be impacted by machine learning, at least to some extent. Low-skilled tasks like administration will greatly be impacted by machine learning. Of course, low-skilled work has more of these tasks than highly skilled jobs. Thus, the impact of automation will be greater in countries that depend on industry and agriculture instead of services, like India, China and Ethiopia. Research by Citigroup and the University of Oxford found that a staggering 85% of jobs in Ethiopia are at risk of being automated compared to 35% in the United Kingdom, based on data by the World Bank Development Report.
Speeding up. The risk of unemployment is also increased due to rapid technological developments. According to futurist Ray Kurzweil, technology develops exponentially. The fast pace of technology, combined with the large percentage of tasks that can be performed by AI, does make one wonder about mass unemployment. But, even if all jobs were to be affected by AI, it does not automatically indicate mass unemployment, as new jobs will be created as well. Take my job as a journalist as an example. Data research and even article writing are already conducted by AI. Quill by Narrative Science makes articles about finance and sports for Associated Press. Other journalism skills are also conducted by AI, such as photo editing, proofreading, and creative writing.
Personally, I would love it if my trusted AI could conduct data research, point out data irregularities and write a basic article about its implications. It would leave me more time to conduct interviews with experts about the findings and investigate the social impacts of these developments. For certain roles, there is enough other work to be done on a higher level, but this is not the case for all jobs. Some skills, including driving and manufacturing products, will turn obsolete with the arrival of self-driving cars and 3D-printing. This would, of course, require low-skilled workers to gain additional skills to prevent their unemployment. Skills that are difficult for machines, such as idea generation, will become more important. But, even these typically ‘human’ skills will be affected by machine learning.
Moving into new territories. We used to think that social skills, for example, could never be conducted by machines. Nursing is one of the areas where social skills are vital. Yet, there is even now a robot called PARO performing some nursing tasks – this cuddly robotic seal is proving effective in relieving patients of stress and aiding recovery. Another area that is not solely human territory any more, is creativity. In photography, for example, ShadowDraw helps you sketch and SmileVector enhances photos easily by turning frowns into smiles. If further areas are affected by AI, the real question to be answered is not how many current jobs will be affected and to what extent, but if enough new job types will be created. Jobs that can only be executed by humans. And linked to this; do Governments realise the threat and act accordingly?
I think with timely and adequate response, we could train enough people in skills that are valuable in the AI-future. This requires a new approach to education; for the most part we should stimulate the education of new professions and stop training people in jobs that will not be around in ten years. If mass unemployment occurs due a lack of adequate governmental action, or because AI moves into further areas, Chace offers a solution to prevent the collapse of the economy: the introduction of a Universal Basic Income to cover our basic needs. At the same time, it would allow us to focus on creating art, play and innovation.
Some people are afraid that basic income will turn us all into couch potatoes, but research in Dauphin, Canada, among 10,000 citizens, showed that only young mothers and teenagers stopped working. If the machines are indeed coming for our jobs, let us hope that by that time a support system is in place to prevent a collapse of our societies.