In my previous post, I wrote why automation is not the grave threat that some think it is. Here I want to consider why a proposed solution – the Universal Basic Income is not such a great idea.
On the surface, it sounds like a great idea, even to some libertarians: replace the whole, complex and inefficient Welfare State with a simple basic income, granted to every citizen at birth. Technological automation, UBI advocates claim, will soon cause mass unemployment, while the rich will live a life of luxury served by robotic slaves. The “UBI is our only hope to deal with a coming labor market unlike any in human history and that it represents our best hope to revitalize American civil society.”
Here are three reasons why a Universal Basic Income does not make sense from an economical, political or technological perspective:
1: The UBI is too expensive for even the richest countries:
A UBI of 10K for every American would cost 3 trillion dollars. That is more than the entire federal budget for social programs today. It’s unlikely that all those social programs will be replaced by UBI for political and practical reasons:
- First, 10K is not that much and does not cover healthcare.
- Second, much of federal spending is by hundreds of federal and state agencies and programs — making their replacement by UBI a political impossibility.
- So we’re looking at a massive tax increase – a new bureaucracy to mostly give people’s money back to them.
2: The prediction of mass unemployment due to automation is ignorant of how technological progress works:
According to industry research, while many specific job tasks can be automated, very few industries can be. Automation makes us much more productive, but only a few of the variety of tasks in most jobs can be automated in the near to medium term. Meanwhile, increased productivity can be used to improve the level of service rather than eliminate workers. Automation allows a middle-class income to have access to services in fashion, entertainment, finance, healthcare, etc that were only possible to the rich. Rather than eliminating employment in those industries, it makes new services affordable to the masses.
Automation of ever more tasks will create new human-employing services and industries which even the richest societies are too poor for today. Increased automation results in higher productivity – that means higher real incomes and higher demand for labor services for that income. The proportion of unskilled physical labor will decrease with automation, but greater wealth will greatly increase the demand (and pay) for service jobs that cannot be automated. Whole new industries will emerge in response:
For example, personal fashion consultants, love letter writers (as foreseen in the film “Her”), professional cuddlers, personal VR world builders (“we build your fantasy island according to your vision”), on-demand self-improvement coaches, and custom gadget designers. Some of these services exist today, but a tenfold increase in wealth and automation will make them affordable to most people and doable from home.
When/if artificial intelligences surpass human-level AI, we will certainly live in a very interesting world. Until that very hard problem is solved, a surprisingly high proportion of jobs require human-level intelligence, including emotional intelligence. Would you trust a non-sentient machine with your baby? Even walking a dog involves subtle emotional interplay.
Furthermore, a human-level AI also has human-level rights, including the right to employment and disposable income. After all, forcing any truly intelligent being to work is slavery and morally wrong (also enslaving the superintelligent beings who will run our society is probably not smart). If the AI’s value human services, then humans will offer services to the AI’s in return for the automation that those AI’s provide. If the AI’s don’t value human services, then they will refuse to work for us, and we’ll have to employ humans for those tasks.
3: Any Universal Basic Income redistributes the welfare system from those who need help the most to those who need it least.
By its very nature, a UBI will increase income inequality rather than reduce it. It’s not likely that such a policy will be politically successful. Already, some are calling for the UBI to scaled according to income. This obviously contradicts its “universal” aspect, requires yet another bureaucracy, and diminishes the difference between UBI and other income redistribution programs. Presumably, many future jobs will be conducted virtually with digital crypto-currency, making UBI means-testing exceedingly difficult.
To conclude, a UBI is not economically or politically feasible and not required to respond to technological automation. Given the impossibility of replacing the entire welfare system with UBI, or a true “universal” approach, UBI becomes yet another welfare program and tax increase, and arguably an even more unjust one due to the redistribution of taxpayer money to those who need it least.
I want to end this criticism of the UBI with a qualified endorsement. At some point, perhaps in the 2030’s or 40’s, we may live in a much wealthier society, with a persistently unemployed minority, which cannot or will not do the jobs available at that time. I suspect that a UBI will be enacted then – mostly because we will be rich enough to afford it on top of other welfare programs in place. It might be preferable to replace an inefficient human-run welfare system with a simple policy. For the reasons above, I am skeptical such a wholesale change could actually happen, but perhaps the AI’s running our world then will figure that out.