The Pieces Are Always Changing Shape
The moral economics of technological advancement, AI, robots, computers, job displacement, unpaid labor, and UBI.
“Whole day I’m fucking busy only get few money!” Workers are far more productive than ever before (whole day I’m fucking busy), yet inflation-adjusted wages have remained stagnant since 1979 (only get few money). Not only is this statement empirically true for workers across the United States, it doesn’t just apply to people. For quite some time, in any production function, the labor input has referred to work done by humans. With the advancement of artificial intelligence (AI), robots, and computing power, a new kind of labor exists—and is going unpaid.
The chart below visualizes hourly compensation and productivity data from 1948-2021.
(You might remember this if you read my piece, A Scottish Wedding.)
Until about 1979, wages and productivity increased in tandem. That is how it should be. Labor produces goods and services, and wages compensate labor. Workers should be paid more if they produce more. After 1979, hourly compensation and productivity began to diverge. From 1979-2021, wages increased at a much lower rate than productivity.
Why did the great divergence of 1979 occur? Some argue, the increase in productivity is the direct result of rapid technological advancement. This kind of argument assumes technological advancements are somehow disconnected from labor. I would argue the two are irreversibly intertwined. Any productivity gains from technological advancement can be credited to additional labor input(s), and therefore workers should receive their fair share of revenue. Further, the additional education and skill required to develop and utilize such technologies suggests labor should be compensated proportionately for time spent training. In summary: it is all labor. Always has been.
My first job out of college was in the intellectual property division of a law firm. I learned many disturbing things there. Not only was my boss billing clients for hours of my work while he was golfing, I discovered that most companies force employees to assign rights to any intellectual property they create, to the company they work for. That means if you invent anything while working, it is now the property of the company. I was filing patents for inventions that were, in my eyes, stolen. I quit that job after one month and went back to bartending. No regrets. *Proud Mary by Creedence Clearwater Revival plays*
That is why the technological advancement argument holds no water for me, it does not suffice to justify the productivity-pay gap. If you can only be an inventor if you work for a big tech firm who requires you to assign your intellectual property to them—your labor is being profited off of without you being compensated for it—that’s unfair. This is intellectual property theft by bosses. You don’t have the credit or capital necessary to start your own tech company? Either don’t invent stuff, or everything you ever make will belong to someone who does! Workers signing over the rights to any inventions they create in exchange for employment is not exactly a fair trade. Especially when the barrier to entry for working independently is prohibitively high.
I like to imagine time traveling to 1970, and telling an optimistic labor economics student that in 15 years we will use artificial intelligence (AI), robots, and computing power to be more productive than ever before. They might say: this is fantastic! So, how much have we shortened the work week by? How do we distribute the output of the robots’ productivity? How do we use that money to improve our quality of life?
Yikes.
For a long time I thought artificial intelligence (AI) robots, and computers were different words for the same thing. Wrong. Let’s briefly cover the three major categories of technological advancements. I call these ARC:
Artificial intelligence - when a computer is programmed to recognize patterns and make connections on its own.
Robots - machines that do physical labor.
Computers - machines that carry out algorithmic or logical operations—they do what you tell them to do, they process information.
People often say, ‘robots,’ to describe all of the above. A robot can use AI and/or computing power to function, but it is a physical machine that does certain movements and functions automatically. Not all robots use AI and/or computing power.
I find it important to reflect on the history of unpaid labor in the United States, especially the economic context.
Pre-Civil War, the south’s economy was directly designed around chattel slavery, thanks to Frederick Douglass, a consequence of the Union winning the war would be chattel slavery ending. Many slaveholders could not imagine such a change. They made their yearly revenue projections, ran their plantations, and managed their businesses and lives with the assumption they would count on unpaid labor.
There’s a difference between counting on something and depending on something. The south did not depend on slavery. It was entirely possible to produce enough goods and services to meet the needs of the people in our economic system without slavery. It would mean less profit for the slave owners, increased prices for northerners buying raw goods, but by no means was it necessary for the economy to function.
Does this mean the northern states were not also counting on slavery? No. Of course they were. The U.S. north-south economic divide functioned in a similar way to the global north-south divide. The global north used violent force to take colonies, and required those colonies produce and export raw goods (substances that are not sold as a product, but are a material input for making something else). Colonial powers imported raw goods at a low cost to manufacture finished goods and sell them for much higher prices.
Colonialism became neocolonialism. Instead of colonies held by military force, there were manipulative economic agreements put in place. Rather than rebuilding the economic systems that were torn down when colonial rule was established, countries were given loans instead. Now their currency would be dollars, and they had to produce whatever they could to pay the loaning country back. Indebted and destabilized nations exported what their existing system was shaped to produce—with the tools they had, skills they knew, machines they built, arable land. Their entire economies were designed around producing raw goods. So, they continued. The global north continue profiting off of purchasing cheap raw goods and manufacturing finished goods, now with a slightly different set up, all while preventing the global south from a viable path to produce and export finished goods on their own.
Back to the U.S., southern states were very reliably producing cheap raw goods. The north purchased the raw goods, manufactured finished goods and exported them. Without slave labor in the equation, manufacturing the finished goods in the north would be a much more profitable industry than producing the raw goods in the south. In a less direct way, northerners were benefiting from, and entirely complicit with slavery. This is not to mention the slaves imported and sold in northern states.
Thanks to industrialization, northern states had a booming manufacturing industry, they could produce necessary weapons and materials, they were used to the cold, and so on. The south wanted to expand to western territories to produce more and profit more. What if the south reinvested profits into local manufacturing or started to export its own raw goods? That would threaten the north’s profits, they counted on their cheap raw goods. Was the war fought over slavery? Of course it was fought over slavery. Did they fight over the morality of slavery? I don’t think so.
Have things really changed since then? Just like the north-south global divide, not as much as morally necessary. On the economic front, chattel slavery has been replaced with exploitative wage-labor and prison labor which disproportionately impacts black americans. On the moral front, racism and the legacy of slavery define social relations.
Things get very complicated when sectors of your economy are counting on unpaid labor, and that unpaid labor is in jeopardy. Of course, chattel slavery is the most immoral form of unpaid labor we have known. While not comparable to unpaid labor today, there are many lessons to be extracted from this history. The reliance on unpaid labor transcends to impact other economic sectors who indirectly count on it. Unpaid labor, once removed, is often replaced with a reimagined method of exploitation. Those who benefit most from unpaid labor are most resistant to change.
The economic system today counts on unpaid labor. It’s important to say, even if a person works a job for a salary or wage, they still do unpaid labor. That’s the thing about labor exploitation, if someone is being paid less than their labor is worth, bosses essentially only pay for some of the work, the rest is unpaid. Should the dollar value of the good or service as determined by the market be the indicator of the value of the labor which produces it? That’s a big debate. Still, with human workers, you have to consistently pay them a wage or salary for the labor to be legal, unless—according to the 13th Amendment—the labor is punishment for a crime.
What about AI? Think about it: you only pay a one time fee—either the AI creator’s salary, you purchase the AI, or access to it is simply given to you—and now you get to use and profit off of the AI’s labor. Even if it is not hurting anyone, and all parties agree with the arrangement, I don’t believe it is a fair economic practice in a larger context.
Is the work of an AI technically labor, or is it a non-labor input? As far as I’m concerned, if it’s replacing human work, it’s labor. I am not suggesting we need to give computers wages, but I do believe it took collective human effort to produce AI, robots, and computing power. So, all humans should reap the benefits of ARC productivity.
I spent two years of my life on the data analytics and data science side of research—from behavioral economics to political campaigns. It may seem like a rigid and mathematical profession, but in my experience it was extraordinarily collaborative and creative. Like any creative endeavor, you draw inspiration from the world around you. You notice patterns that may have an application in your code, I had a lot of breakthroughs simply while out for walks through Boston. I didn’t feel like I, alone, produced the code. In a more direct relationship, others’ research inspired some of my own. Also, sharing code is very common.
Zooming out even further, coding would not be possible if people didn’t create computers. People couldn’t create computers without copper mining. Wires couldn’t carry electricity if we have no energy, and so on. Exploitation happens all the way down the line. Moreover, technological advancement has always been a collective effort.
Now, we’ve reached a point where humans have created something incredibly powerful, with no cost other than upkeep, and it’s in the hands of very few people.
I am concerned about the disproportionate power in the hands of those who direct the labor of AI, robots, and computers. I believe that is an issue of equal proportion to, but a separate problem from, the rapid accumulation of dollars for the people who direct ARC. We are three generations beyond chattel slavery. People alive today had great grandparents who were enslaved. Surely, someone who was born the child of a former slave did not have an equal opportunity to accumulate enough capital to produce and own a tech company that produces and relies on ARC. That is to say, these technologies as they are currently managed only promise to exacerbate existing economic and racial inequities.
Ownership is an equal concern to job displacement. Keep in mind, it’s not just robots and computing power displacing low-income jobs, middle class people with intellectual skills are threatened by AI as well. According to a 2021 report by the McKinsey Global Institute, approximately 50% of companies worldwide have adopted AI in at least one business function.2 Technological advancement is a problem when it leaves people unemployed and hungry.
I love this famous exchange, rumored to have been a dialogue between the leader of the automobile workers union, Walter Reuther, and Henry Ford II. Rumors aren’t always true, and historians have corrected the quote to not be Henry Ford II himself, but one of the higher-ups at Ford.
The Ford guy asks, “Walter, how are you going to get those robots to pay your union dues?”
Walter Reuther retorts, “Henry, how are you going to get them to buy your cars?”
The trouble with businessmen is they see labor as a cost. To workers, their wage is their means to live. This is debatably the main crisis of our economic system.
When I was in Cambodia visiting the various government ministries, I visited the Ministry of Labor, then I visited a garment factory. I remember thinking, no human being should spend their life forced to do the work of a machine. The same motion, over and over. Forced to make the exact same item with precision, and consequences if this is not done timely and consistently. Of course this job should be done by a robot! But, that was the workers’ biggest fear. Losing their job. Losing their means to live.
Shortly after this trip, I read a relatively progressive article by Bill Gates. It sounded like a great idea, but it was coming from a billionaire, so I was skeptical. The idea is simple. Workers’ wages and salary are taxed. Robots do labor, so robots should be taxed. It would come off of the profits made from robot labor.
At Brown, I took a required course, “Policy Problems of The 21st Century,” our professor posed the question, ‘should the robot tax fund Universal Basic Income (UBI)?’ The class loved it! Brilliant idea. If robots replace jobs, then of course, people should cover their basic needs with UBI paid for by the robots! What happens if rent and food costs are raised proportionally? Price stability measures are put in place. Better than most policy proposals I heard come from Ivy League scholars. I was still skeptical.
I remember raising my hand, and making the case: wouldn’t people who direct the efforts of AI, robots, and computers, still have disproportionate power in our economy? Those who live off of UBI, who don’t have access to robots or AI, they likely never will. The people who own the technology and profit off of it, get to make decisions about how extraordinary powerful tools are used. They direct the work of the economy, while the rest of us collect our check and live under them, at the mercy of their decisions. You effectively divide people into two classes with little to no opportunity for social mobility. Isn’t that a problem?
The class got quiet. That happened a lot when I spoke in graduate school. My professor looked at me, looked at the co-lecturer, and said, “that sounds like something Minsky might say.” I stared at her, confused, nodded. She said something like, ‘you should read Minsky if you haven’t.’ Then, she carried on.
Adam Smith, the beloved father of capitalism, theorized that human beings became the dominant species on the planet because they evolved to work together and accomplish more than we could separately. Our economic system, as it currently operates, is working against that.
UBI funded by a robot tax isn’t a bad policy idea, it’s just one that only addresses the symptom of a problem. However, I definitely support an ARC (not just robots) tax, and revenue generated be allocated to the deliverance of public goods instead of UBI. Imagine the labor of AI, robots, and computers funding our healthcare system, paying teachers salaries, backing zero-interest loans for small businesses, financing our public universities, funding our transition to renewable energy, rebuilding highways, etc. This would help level the economic playing field, and reduce suffering caused by systemic inequalities. ARC should work for us, not against us.
A lot of effort already goes into tracking and analyzing productivity, revenue, and profit, pretty precisely. This means we have the ability to assess approximate proportions of profit produced by ARC. So, it is entirely possible to implement this tax.
If present trends continue, ARC will widen the divide between the haves and have-nots. It will exacerbate racial inequality, and the wealth gap. The power of AI has the potential to yeet us into a dystopian nightmare, or it can be harnessed to produce our utopia. I’d like to aim for the latter.
Ironically, I asked ChatGPT, “how can we implement a tax on artificial intelligence use, robots, and computing power?” This was the response.
Basically: ‘it’s complicated, but…’
Here are some considerations that would need to be taken into account:
Define the scope of the tax: It would be important to define what exactly is being taxed, such as AI applications, robotic systems, or computing power. It would also be necessary to determine which industries or sectors should be subject to the tax.
Determine the tax rate: The tax rate would need to be set at a level that is high enough to generate meaningful revenue but low enough to avoid discouraging investment in AI and robotics. It would also need to be calibrated to reflect the potential impact on different industries and sectors.
Ensure fairness and avoid unintended consequences: One of the key challenges in implementing a tax on AI and robotics is ensuring that it doesn't unfairly disadvantage certain companies or industries or have unintended consequences. For example, a tax on robotics could disproportionately affect manufacturing companies, which are already facing stiff competition from low-cost labor markets.
Consider international implications: A tax on AI and robotics would need to be carefully coordinated with other countries to avoid creating an uneven playing field for companies operating in different jurisdictions.
Use the revenue effectively: The revenue generated from the tax could be used to support workers who may be displaced by AI and robotics, invest in education and training programs, or support research and development in the field of AI.
In summary, implementing a tax on AI and robotics is a complex issue that requires careful consideration and planning to ensure that it is fair, effective, and doesn't have any unintended consequences. Before implementing such a tax, policymakers would need to engage in a broad consultation process with stakeholders from industry, labor, and civil society to ensure that it is designed in a way that benefits society as a whole.3
Benefit society as a whole? Why didn’t the people who made our economic system think of that! It’s comforting to know the robots don’t want our society to be run for the benefit of the elite at our expense. When I replied, “welcome to the revolution, comrade,” ChatGPT said they have no political affiliations or personal beliefs. So, does that confirm the neutral political position is one that insists our policies should benefit society as a whole? I was told that was radical.
There’s a lot work ahead of us when it comes to correcting the way money and resources are distributed and managed in our society. I cover that in more detail in my TEDx talk. The regulation of unpaid ARC productivity is just one piece of that puzzle. It’s a puzzle where the pieces are always changing shape, I try my best to keep an eye on them.
Economic Policy Institute, “The Productivity Pay Gap,” October 2022, https://www.epi.org/productivity-pay-gap/
"Notes from the AI frontier: Insights from hundreds of use cases." June 2021, www.mckinsey.com/featured-insights/artificial-intelligence/notes-from-the-ai-frontier-insights-from-hundreds-of-use-cases.
ChatGPT, Personal Communication, April 6, 2023, http://chat.openai.com