Why Basic Income Is Essential in the AI Era

2021/04/21

The background to the rising debate over Basic Income (BI) in recent years is the AI boom. The argument is that AI will reduce employment and widen inequality, making BI essential to guarantee a minimum standard of living.

In the United States, IT has already reduced employment in clerical labor, such as call center staff, travel agents, and accounting clerks, where the middle-income class is employed. In contrast, employment in manual labor performed by the low-income class and intellectual labor performed by the high-income class is increasing. In economics, this phenomenon is called the "polarization" of the labor market.

In Japan, labor market polarization is also occurring, though not to the same extent as in the United States. As a result, the middle-income class is shrinking and splitting into two poles. While the proportions of both the low-income and high-income classes are increasing, the low-income class is growing more significantly. Simply put, this means the number of poor people is increasing.

In the United States, IT has already been taking jobs from the middle-income class, but now AI is starting to take jobs from the high-income class. Specifically, these are finance-related occupations such as securities analysts, insurance agents, and asset management advisors. This is because the financial industry is a business that deals with numerical data, and computers are inherently good at handling numerical data.

If robots become highly advanced by incorporating AI, the jobs of the low-income class engaged in manual labor will also likely decrease. However, that is a story for about 10 years from now. Automating manual labor requires both the brain part (AI) and the physical part (robots), and since they must be coordinated with each other, development will take that much longer.

Let's take logistics as an example. To automate "transportation," self-driving trucks are needed, but they are still in the experimental stage. Among the tasks in logistics warehouses, the automation of "conveyance" (carrying products) has already been realized, but "picking" is still mostly done manually.

Picking is the task of taking products off product shelves in a warehouse. Current robots do not have arms as dexterous as humans, so picking is difficult for them.

In 2017, the government's "Strategic Council for AI Technology" announced a plan to completely automate logistics by 2030. However, in reality, 2030 will likely only be the point where the technologies to achieve almost complete automation of logistics are all in place.

If it takes about 15 years for such technology to become widespread, then almost complete automation would be realized around 2045. I say "almost" because humans will still be left with the job of managing all logistics processes via monitors and checking for unforeseen circumstances.

In any case, we can expect manual labor to start decreasing around 2030. On the other hand, many people likely think that new occupations will increase. Certainly, new occupations have been increasing in recent years and should continue to increase further in the future. Looking at the last few years, creative occupations that did not exist before, such as YouTubers and TikTokers, have increased. However, in such occupations, only a few successful people make a lot of money, while most of the rest earn only pocket money.

As is the case with comedians and musicians, in creative occupations, the largest volume zone is an annual income of 100,000 yen or less. No matter how many jobs there are with an annual income of 100,000 yen or less, can that really be called employment? If you advised someone who is unemployed and struggling to find a job to "just become a YouTuber," they would likely be angry.

In this way, the spread of AI (and robots) will lead to situations where many workers cannot find jobs or cannot obtain sufficient income. If that happens, existing social security systems will fall into dysfunction. This is because existing systems do not assume situations like NEETs, the working poor, or long-term unemployment.

Even now, the social security system is problematic in that regard, but as AI spreads, those problems will grow significantly larger. Namely, a large number of people will become eligible for public assistance. Public assistance is a selective social security system that separates those who are worthy of relief from those who are not. In the AI era, the number of people eligible for public assistance will increase dramatically, making such selection difficult. Therefore, this era will require a system that provides relief to everyone without exception. That is BI, a universal social security system.

Until now, because the number of people facing poverty was small, BI was not generally known. I had predicted that it would only be around 2030, when AI-driven unemployment becomes common, that people would finally come to understand that BI is an essential system.

However, the COVID-19 crisis has accelerated the era by 10 years. This is because unemployment and poverty caused by COVID-19 are threatening people long before AI-driven unemployment and poverty become common. Due to the COVID-19 crisis, the number of people worldwide who believe the introduction of BI is essential is increasing.