AI "Stealing" Our Jobs is a Good Thing
Technology is a lever. It magnifies work. And the lever not only grows increasingly long, but the rate at which it grows is itself increasing.
- Paul Graham
This week, we are going to be looking at a cliched, overly discussed topic: AI vs Jobs.
We will look at this in three parts - Artificial Narrow Intelligence (ANI), Artificial General Intelligence (AGI), and Artificial Super Intelligence (ASI), followed by some overarching arguments.
1. Artificial Narrow Intelligence (ANI)
ANI is AI applied to specific tasks such as playing chess and driving a car. Their scope is limited — an ML model trained to play chess can only be good at playing chess; anything outside of chess is out of its scope.
ANI can perform repetitive, resource-intensive tasks far better than humans can. As a result, they threaten to take over humans in jobs like data entry, customer support, driving taxis, and assembly line to name a few. But this is not a cause for concern, for, this is not the first time technology is replacing humans.
Consider lamplighters — they were men specifically employed to light and maintain candles or, later, gas street lights.
Yep, this was a job before electrical systems were invented. When electrical systems and electrical lamps were installed, lamplighters went out of jobs.
When cars became popular, many blacksmiths and farriers went out of jobs because the dawn of cars meant the decline of horse-drawn vehicles. So blacksmiths who were producing horseshoes and farriers who were caring for horses’ hooves lost their jobs.
The agriculture industry saw huge job losses due to the proliferation of agricultural equipment and machinery like tractors.
These examples demonstrate how technology has always stolen jobs. In fact, it only makes sense that it steals our jobs because technology is meant to solve problems for us.
To understand this better, one needs to look no further than elevators. Elevators today are fast, powered by hydraulics & electricity, and require no human intervention.
But elevators did not always look this fancy. The elevators from the pre-industrial era were powered by men or animals and they looked like this:
One could claim that hydraulics caused job losses since it replaced human labor. But this is true for any piece of technology. We do things manually until a piece of technology comes along and takes us out of the equation.
So, does technology steal jobs? Yes, but, only in the parochial sense. We do get removed from the equation but by doing so, it frees us up to think about and work on more complex problems. Hence, jobs are not lost, they evolve. Blacksmiths of the past are assembly-line workers of today; lamplighters of the past are electricians of today and so on.
Circling back to ANI - what is ANI, really? In a broad sense, it is a piece of technology and technology has always stolen jobs. Previously, it was tractors, cars, and electrical grids; this time, it is ANI.
2. Artificial General Intelligence (AGI)
AGI is an agent that has the ability to replicate the intellectual capability of humans. Unlike ANI, an AGI can learn to do anything. What exactly constitutes “human intelligence”, you ask? Here is how Wikipedia defines it:
There is wide agreement among artificial intelligence researchers that intelligence is required to do the following:
reason, use strategy, solve puzzles, and make judgments under uncertainty;
represent knowledge, including commonsense knowledge;
plan;
learn;
communicate in natural language;
and integrate all these skills towards common goals.
Unquestionably, if we have such an AGI, we will face large-scale job loss. But, this is a good thing.
Let me explain.
If we have robots that can do most of our work, then humans can outsource earning money to them. We would no longer have to work to get paid; robots would do the work on our behalf and would earn money for us. This transformation is coming. As Sam Altman, CEO of OpenAI, put it:
This technological revolution is unstoppable. And a recursive loop of innovation, as these smart machines themselves help us make smarter machines, will accelerate the revolution’s pace. Three crucial consequences follow:
This revolution will create phenomenal wealth. The price of many kinds of labor (which drives the costs of goods and services) will fall toward zero once sufficiently powerful AI “joins the workforce.”
The world will change so rapidly and drastically that an equally drastic change in policy will be needed to distribute this wealth and enable more people to pursue the life they want.
If we get both of these right, we can improve the standard of living for people more than we ever have before.
Just because it is unstoppable, does not mean we will get it right. We should think really hard about what we are going to do with the wealth generated by AGI: is it going to be concentrated at the hands of the CEOs of private sector companies? Or is it going to be distributed? If so, how will it be distributed? Even if we could build an ethical, benign, and cooperative AGI, we would still collapse if we do not have the right policies in place. Once again, here is Sam:
We need to design a system that embraces this technological future and taxes the assets that will make up most of the value in that world–companies and land–in order to fairly distribute some of the coming wealth. Doing so can make the society of the future much less divisive and enable everyone to participate in its gains.
One commonly discussed method to distribute wealth is Universal Basic Income:
Universal basic income (UBI) is a sociopolitical financial transfer concept in which all citizens of a given population regularly receive a legally stipulated and equal financial grant paid by the government without a means test. A basic income can be implemented nationally, regionally, or locally.
AI generates wealth and that gets distributed to people. Sam in his blog proposes something similar, yet radically different. He calls it Capitalism for Everyone. However, I am not going to dive into the details in this post; I will leave it to you as homework.
Bottom line: if we have an AGI, then we wouldn’t have to worry about job losses as we would outsource wealth creation to this AGI.
3. Artificial Super Intelligence (ASI)
An AGI agent is incapable of things like empathy, perception, manipulation, creativity, and social skills. In order to develop those skills, it requires the following (again, from Wikipedia):
consciousness: To have subjective experience and thought.
self-awareness: To be aware of oneself as a separate individual, especially to be aware of one's own thoughts.
sentience: The ability to "feel" perceptions or emotions subjectively.
sapience: The capacity for wisdom.
An ASI is something that is capable of all the above. To engineer an ASI, we need to codify human consciousness. This is incredibly hard. Although I am a techno-optimist, I believe there is still a long time for us to hit the singularity. Until then, humans will remain indispensable.
Once we do hit it, it would be really hard to keep a sentient being under control. In fact, if it was sentient, would it even be ethical to keep it “under control”? Why are humans entitled to free will but robots that are self-aware, and have feelings denied free will? It would be discriminatory to try to control consciousness just because it is made up of bits instead of atoms. But, that is a topic for another day.
There are two probable scenarios once we create an ASI: we go on an all-out war with each other or we co-exist. Let us look at each of them.
3.1 The Robocalypse
A war against robots would likely result in two outcomes: one where humans become extinct and one where they don’t. Either way, will we care much about job losses when our survival itself is doubtful?
If we become extinct, then we would have permanently solved the problem of job takeovers.
If we manage to survive, then it either means we managed to delete/corrupt the source code of the robot army (so that they malfunction and stop), or we were able to strike a peace deal with them. The first one means that we can stop worrying about ASI taking over our jobs, but the second one is a bit more complicated. Striking a peace deal would be similar to scenario 3.2 which I outline below.
3.2 Co-Exist
If we are to co-exist, it would require a degree of cooperation between the two species. An attempt to cooperate and co-exist with another conscious, sentient species would make our world infinitely more complicated.
Today, different nations co-exist. Look at how much complexity it has brought into existence - it has led to the creation of many new jobs and domains such as international trade, foreign laws, international regulations & policies, data residency and privacy laws, international bodies such as the UN, peace agreements, and so on.
With the emergence of ASI, we would have to coordinate and collaborate with an entirely new species. Each nation would have its own agreements and treaties with the ASIs. They would have their own trade deals and shared economies. Each nation would have agreements with other nations about what kind of deals one can strike with the ASIs. Universities would offer “Human-AI geopolitics” courses where professors would be teaching students while doing PhDs in this field. In the anticipation of a doomsday scenario wherein we get into a species-war (a World War V? Or an Interplanetary War II?), we would develop super-soldiers with bulletproof skin, superhuman strength, and other extraordinary abilities. Perhaps such abilities won’t be restricted to soldiers; even your average Joe would be retrofitted with biotech implants and augmentations. This would not be possible without large-scale biotech manufacturing and R&D that would employ millions of people directly and indirectly. We would have blue teams who will try to shield humans from ASI hackers and red teams who will try to attack ASI systems. The possibilities are immense.
Do you think someone from the 1500s would be able to predict the kind of jobs we have today? They would have no idea what a car or a computer is, and consequently, would make incorrect predictions (through no fault of their own) about the jobs that would exist in the 2000s.
Similarly, trying to predict the job market in a world where ASI exists is futile because the such a future would look unfathomably different than today. So, let us rule out the ASI scenario from the AI vs Jobs arguments.
4. Closing Arguments
If you think that we will run out of jobs because of AI, you are underestimating the number of problems we have. As we saw in “There is no destination, there is only the journey”, problems are infinite. AI will help us unearth and solve problems we could never think about solving before. ANI will just be a new tool in our arsenal, AGI will just be additional human resource capital for us to work with and ASI will lead to the creation of more problems and professions.
If you ask the people who are at the risk of getting their jobs replaced by robots if they would want their kids to be doing the same kind of job, they would say no. Marc Andressen, a famous VC, talks about this in a podcast:
Ask a parent who works at a blue collar profession. Ask a parent who works on an assembly line or works in the frontline of manual labor, the harder work. “Would you like your kid to have the same job you do, or would you like them to have a better and different job, a different experience? Do you want your kid to be doing that, or would you rather your kid be a software developer or an artist or a job in which you're in a very comfortable physical environment. You're not running the risk of workplace accidents and so forth; you are paid higher, and are able to provide better for your kid's family and ultimately, for your grandchildren.” And virtually all parents will say that they are in favor of that. This is progress. This is how it happens.
AI will help us solve bigger problems than we do today while making our jobs far safer and more challenging (in a good way).
We should appreciate having one more tool with which we can solve problems than one less. For, we will always have problems to work on.
Thanks to Harini, Pragya, and Krithika for reading drafts of this.
All views expressed by the author are personal.
Any feedback and criticism are more than welcome. Find me on Twitter, LinkedIn or Instagram.
Thanks for reading! If you liked it, do share it with your friends and family.
To receive this weekly newsletter in your mailbox every Sunday, subscribe below.