One of the topics that is inevitably tied to the big digital transformation themes like AI, machine learning, robotics and digital transformation in general is the universal basic income (let’s call it UBI from here). Like it or not, it’s worthwhile considering which tools, mechanisms and concepts need to be adapted or created when this massive transformation will roll over us. I believe it would be naive not to assume that our political economy, and governance will desperately need new design patterns as we enter this new phase of the digital revolution.
There’s different ways to position or regard the concept of UBI. Ranging from:
- liberating people from poverty-level jobs,
- to creating a cushion for the weaker or less or not at all employable who are not catching up with the innovation and economic growth delivering market forces,
- to UBI being a dedicated response to an era in which human labor is less and less viable and competitive in an all digitally transformed and automated production and services system,
- to allowing people to devote time to creative, artistic, charitable exploits.
A look at the underlying drivers
I think except for the latter, they are all somehow related, or respectively are spurred by the same drivers. And these are the drivers that in my opinion determine the times we live in and that – at accelerating speed – will determine future times:
- growing world population and thus more people competing for labor and jobs
- rapid and accelerating technological innovation
- which allows algorithms and machines to eat deeper into previously “only human labor feasible” tasks and activities
- thus ever-increasing need for humans to qualify for cognitively advanced tasks, alongside respective education/training/qualification/skills (“the ticket to the knowledge economy”)
Even without number 1, numbers 2 to 4 form a repetitive and self-supporting cycle leading to more labor (blue and white-collar) taken over by non-human actors and thus less labor/jobs left for humans. Unless of course totally new jobs and tasks are created for a growing number of people.
Some claims against UBI drive me crazy
There’s a couple of articles, opinions and judgments I have stumbled upon recently on UBI that kind of started to make my blood boil a little bit. What strikes me, is the apparent degree of “not being informed” or at the least “narrow vision” that can be observed. I get the impression that the stronger the opinion against UBI, the less informed the knowledge and consideration basis seems to be.
“Mankind is hard-wired to work. We gain satisfaction from it. It gives us a sense of identity, purpose and belonging … we should not be trying to create a world in which most people do not feel the need to work.”
(my quick stance: UBI is not about putting people out of work but exactly the other way round – it is an answer to a broad trend of people being pushed out of work by technological progress)
(my quick stance: yeah, right… when well-paying jobs are on the decline for clearly relatable reasons, let’s just still proclaim we want those jobs!)
(my quick stance: this is absolutely right… but doesn’t give an answer to the trend of less and less jobs that provide this chance for receiving dignity)
(my quick stance: we don’t need human-level artificial intelligence or artificial general intelligence to severely disrupt our labor market… the artificial narrow intelligence examples already observable today are scary and impactful enough)
(my quick stance: again… yes, it would be great if there was something left for humans now that machines and software are eating into complex manual labor and higher level cognitive tasks, but what field is left when manual and cognitive labor are taken and how likely is it that such jobs will occur in the billions?)
And I am not even talking about opinions like “if you give people UBI then they will be lazy and never work again”. Remember, we’re not talking about the world in 2017, but most likely the world in 2030 and onwards. And then it will not be about people not pursuing a job because of UBI, but UBI should be there because they cannot pursue a job in the first place.
Do we need artificial super-intelligence before we need to be scared of automation?
Before I move on to elaborate on the argument, that things will just work out like they did in the past (“old jobs go, new jobs will come”) I would like to take a look at the “AI is not smart enough to justify being scared”. It is not a strong argument that certain AI algorithms are only good at one thing but not able to generally compete with human intelligence or capabilities.
Critics say that all current AI progress is respectively only centered on one field or domain. AlphaGo can play Go, but not peel carrots. So what? Then there is one system for mastering Go and one system for peeling carrots or flipping burgers. Not one single system needs to master everything. As long as there is an AI solution mastering tax law and accounting, another one mastering medical diagnosis, and another one manufacturing a car, and another one driving that car, we’re still in deep shit. We don’t need strong AI, not articial super intelligence, no intelligence explosion, no singularity to start sweating when thinking about automation and job replacement.
The severity of the discussion should get even clearer when considering concepts like reinforcement learning in the machine learning discipline. They show us where we are heading. This is about algorithms learning what to do from scratch, without human intervention, surpassing human capabilities in a certain field within hours.
Will everything be alright – just like we have seen in past times?
You often hear the argument that technology is both creating and destroying jobs and that this has always happened. “Just look what happened in the past.” Agricultural and handicraft trade jobs have been destroyed and replaced by industrial jobs. Industrial jobs have been destroyed and replaced by service and knowledge worker jobs. The carriage driver was replaced by the car. But then there were new jobs like mechatronics engineers or automotive engineers.
But what comes next? Yes, it sounds like a logical development that when we replaced physical labor by machines, people would find new labor in more intellectually focused professions. But what when you replace intellectual labor conducted by humans with software logic and algorithms? What’s next then?
To me it is simply an uninformed contribution to the discussion to argue that exactly this chain of events will continue and that there will be something next. Talking to people about this topic, you also find ingredients to the discussion like “we humans cannot allow that”, “it is inhumane to allow for machines to replace human labor” or just the simple claim for “we must create new jobs”. Like we have IT-administrators, and web developers, and app developers today. All of them jobs that didn’t exist 25 years ago. But how many people work as web designers? Isn’t it the truth that the new jobs in technology that have been created are by far lower than those numbers of jobs that used to exist in manufacturing and classic industry focused professions?
I did some research and dug up these numbers:
- Google/Alphabet, Facebook and Apple
- 2016 revenues: ca. $332bn
- 2016 year end market cap: ca. $1,500bn
- 2016 year end employees: ca. 205k
- General Motors, Ford and Fiat/Chrysler
- 2016 revenues: ca. $428bn
- 2016 year end market cap: ca. $120bn
- 2016 year end employees: ca. 660k
Tech companies employ significantly fewer people than traditional industry players while playing in a totally different ballpark when it comes to market capitalization. And it’s only a matter of time when those three tech players will surpass the industry players in terms of revenue. In terms of profit they are already in another dimension today.
New general purpose technologies will turn our labor market upside down
So, looking at the technologies that will further drive progress in tech and increase pressure on automation and replacing human labor, AI and robotics have huge potential. I believe artificial intelligence and robotics that is AI-powered are general-purpose technologies, like e.g. electricity. We’re seeing a definite ramp up and acceleration of progress and utilization of such technologies in our economy right now. And even – as can in fact be noticed – when there is a lower than expected productivity ramp up, this in my view is likely an up front condition as companies invest in these technologies without necessarily seeing a direct return in the short-term. But this will come later. And it will come with force. It is hard to imagine what kind of jobs there will be for humans in the very long run once machines and robots have reached levels where they easily surpass human labor in the same field both with regards to time and costs.
Three phases in the “human – machine” mix until humans are out
In the short to medium term I foresee the following development.
First, intelligent machines and algorithms will augment human labor, so that its productivity will be significantly increased. In the combination of human and machine labor, the human value in the overall contribution level will be higher than that of the machine.
Secondly, intelligent machines and algorithms will make significant progress in their capabilities so that human contribution and interaction will be needed to close the last remaining gaps. Think “intuition” and “common sense”. It’s that ingredient to decision-making and supervision that is hard to describe and to grasp, but that makes all the difference. In this sense, humans will have the role of supervising, but altogether the share in the overall value contribution decreases significantly.
And lastly, even this last resort will be covered by even more capable machines and software algorithms. Leaving the human out of the equation. In this stage, human involvement would be counterproductive, probably even massively inferior. This will go to a degree where it is difficult to grasp for humans how those algorithms actually work, both from an intellectual point of view and from a speed of processing point of view.
And then it boils down to the question how we envision human participation in income and wealth distribution in such a scenario? Concepts like a universal basic income seem like a pretty reasonable approach to me.
Our thinking must change from past environmental conditions to those of the future
Coming back to claims like “but humans need work for dignity”, this to me feels like thinking trapped in past conditions. Contributors should consider that we are not talking about today but about a future that has ever more accelerating progress of technological development. It is crucial to anticipate an upcoming development, whose mechanics are based on exponential development and growth and that is in itself further fueled by the law of accelerating returns.
Yes, it is a valid and comprehensible claim to provide enough work for all people. To have a fair (or somewhat fair) share in wealth distribution, in finding meaning in life and gain satisfaction from a professional career.
Figuring out UBI is challenging – but there is hardly an alternative
But the likelihood that this will be possible is getting smaller and smaller. Therefore it is a necessity to explore how a universal basic income could work. I don’t know how and if it will work. But elaborating on it to me seems much smarter than to claim that we should not give up against automation and the machines and simply invent new jobs. Yes, there will be jobs that we have no imagination of so far. But I doubt those will keep billions of people busy and employed.
Don’t get me wrong. It is extremely challenging to figure out a working UBI scheme. It is extremely challenging to figure out the math and financials. However, I believe it is more worthwhile spending our time and brain power on tackling this, than to argue that we “simply need more and new and better paying jobs”. I don’t see those coming in the quantities needed to sustain our current way of working and living. I think technological progress is going destroy the labor market as we know it, and it’s going to create a desperate need to find solutions in order to provide social stability.
Read on LinkedIn Pulse
Read on Medium