Which One Are You? (AI)
There's an interesting active debate going on in the world of all things AI. This current debate, though, reminds me of the following joke.
There are three kinds of people in this world - those that can count and those that can't.
Which One Are You? (AI)
There are three types of people participating in this AI debate, those that are Pro-AI and those that aren't. Among those that aren't, some fear AI while others are just in it for the long run.
I acknowledge that there are AI advances, like in healthcare and scientific developments. There are also questions posed that are simple and obvious for humans to answer correctly while AI answers incorrectly (like telling someone it's better to drive their car with a flat tire to a garage that is 50 meters away) and this is referred to as its jagged edge. Also, AI is not going away. It's worth it to study ways to ensure favorable outcomes while avoiding negative ones. This is not an article claiming AI is either bad or good.
This piece is an attempt to express out loud what I believe the 'long run' actually means. It's a period of real time experienced by real people. History has something it can teach us about the 'long run' if we're willing to look carefully.
My input into this debate focuses on the most common points of view. First, some observe that incorporating new and innovative technology often initiates some noticeable number of job losses in certain areas. Second, some observe that technology eventually creates more jobs than those that were initially lost while also often improving the overall lives of most people in the long run. History provides several examples showing that this second observation is justifiable. However, it's also important to include a third observation about what happens during this long run.
AI is an innovative fast moving technology and a few corporations claim some recent job losses were caused by incorporating AI into their work environment. After all, AI models are designed to replace tasks performed by human workers. Speculating about what could happen during the long run is something I claim should be part of the debate. AI will either increase or decrease the number of jobs in the long run. We don't know yet. AI will either make the living conditions better or worse in the long run. We don't know yet. There's not enough data available to analyze so we can't claim we know with any certainty what the future will be. This uncertainty means we do not know, during the long run while we wait for more data to show us that everything is getting better, if some severe negative events might happen.
Typically the Pro-AI forces, claiming positive economic outcomes are likely in our future, cite previous positive results from historical periods where there were technology improvements that eventually improved lives. One example is the Industrial Revolution (a lengthy period of time aptly named by historians). Although it is true that the innovations and efficiencies emerging from the Industrial Revolution did eventually improve most lives, it is also worth noting that it took a very long time, and there were events, during the long run, that negatively affected workers and their families.
If you start looking from the beginning of the Industrial Revolution (let's say mid 18th century in Great Britain) and then examine the events that workers experienced during the long run of the Industrial Revolution period (say until World War I) you will find there are events that should be studied because they reveal lessons we can all learn from. To understand what the long run can actually look like for workers, it's worth spending some time looking at workers during the Industrial Revolution because this is an example where the long run spans the entire lifetime of many workers.
The following is my abbreviated and simplified history of the industrial revolution. At the beginning of the Industrial Revolution, 20% of the British population lived urban lives. The innovations that occurred over time allowed for the mechanization of farming. This mechanization forced many rural residents to move to urban settings as the number of jobs available in factories was increasing, while at the same time, there were fewer rural farming jobs available. About 100 years after the beginning of the Industrial Revolution one finds that 50% of the British population now lived urban lives. Some conditions workers experienced during the Industrial Revolution are best described as hellish, dismal, dangerous, and fast paced with very long hours every week. Children also worked in factories and too many developed disabilities that carried into adulthood. Services offering compensation for injuries were rarely available as the kinds of regulations needed largely did not exist. By the end of the Industrial Revolution more than one generation of workers suffered and basically lived miserable and impoverished lives. Friedrich Engels (German philosopher) wrote the book "The Condition of the Working Class in England" which was published in 1845.
Friedrich Engels (Wikipedia)
"It presented a wholly unflattering account of the 'possessing class' and its role in a competitive economic system, emphasizing the worst cases of poverty and degradation. Engels detailed the squalor of the slums, the brutal exploitation in the factories, the "social murder" committed by the bourgeoisie, and the systematic spatial segregation of the city that hid this misery from view." (Wikipedia)
This period is sometimes called "Engels' Pause" which is a historical term coined by economic historian Robert C. Allen who validated Engel's harsh findings. This is a period of time during which productivity and output rose steadily while workers' real wages and living conditions stagnated or declined.
"Before the Industrial Revolution, agricultural needs dictated pace of life. Farm workers faced seasonal, dawn-to-dusk labor dictated by daylight and weather. Conversely, factory workers during the Industrial Revolution operated on strict, mechanized schedules inside crowded, dangerous urban facilities." (National Geographic Society)
No claims can be made that rural farm workers never experienced any misery or injuries.
An examination of historical findings reveals that it's reasonable to claim that lessons can be learned by examining the events that show living conditions deteriorated significantly as workers moved to urban factory working conditions. Conditions did not improve on their own. The innovations of the Industrial Revolution created improved efficiencies and productivity but did not initiate improvements to working conditions. It took rioting and striking workers, and most importantly, the creation of labor unions before the demands for improved working conditions were heard and acted on. These working conditions did not meaningfully improve until workers struggled through slow but steady growth of union membership, that eventually resulted in labor becoming powerful enough to force factory owners to accept their demands. The United States followed a similar pattern as the British. In 1800, roughly 90% of Americans worked in agricultural related fields. By 1900 that percentage had fallen to about 40%, and today it's under 2%. The transition reshaped the country, and for many workers and families, the long run was their entire life.
For me, the question is not just did lives improve in the long run (because they did), it's also about this very long period where too many workers paid a significant toll. A lesson I take away is that perhaps there is a better way to approach unknown futures when there's concerns about possible serious health conditions and negative economic outcomes affecting too many people. It is clear that just leaving everything up to the current wealthy few ('possessing class'), who are the main beneficiaries of the improved efficiencies and productivity gains, will mean that few if any solutions addressing the immediate needs of workers and communities will be proposed. Workers also make considerable productivity and efficiency contributions but receive little additional compensation. They are simply told that life is not fair and eventually it will get better in the long run.
The main takeaways from studying the working and living conditions of factory workers during the long run of the Industrial Revolution are the following:
These takeaways influence how I answer questions about whether the lives of workers improved over the long run. My answer takes into account the amount of time and level of misery workers and communities endured during this period of time. Similarly, answers to questions about whether technology in general eventually improves people's lives, must include considerations about the length of time and level of discomfort experienced. There is also a need to include, from the beginning, those people needed to help define what outcomes are worthy of observation, how to evaluate these observations, and what actions are needed to prevent the emergence of unacceptable negative outcomes.
One example is the Internet revolution that started in the 1990s and was firmly in place over a 10 - 20 year span. New jobs emerged quickly. Although certain industries, like retail, print media, and travel agencies never completely recovered, this is largely an example of technology creating positive results without causing a large population of workers to experience an unacceptable level of negativity in their lives. Most historians and economists do not raise serious concerns about this period because the long run was a much shorter time period, new jobs were created quickly, and the new jobs paid well. The question this raises for me is, how will AI affect this Internet revolution in the long run?
A contrasting comparative example, with workers, families, and communities experiencing different outcomes, is available by studying recent manufacturing automation and globalization. The number of manufacturing workers was significantly reduced, communities were harmed, and even after an ongoing 50 year long run time period many have not yet fully recovered. It doesn't feel right to me to just say that if we just wait longer, the improvements will eventually arrive. There is often more that should be said in the short term. Will AI offer any help to these workers and communities that were adversely affected in the long run?
If You're So Sure - Prove It? (AI)
AI is a current technology focus that is rapidly delivering increases in innovation and worker productivity. The wealthy few are a key faction that is promoting rapid introduction into the workforce. Those holding the point of view that AI will, over the long run, improve most, if not all, lives is a position that cannot be disproved. It also cannot be proved. We need a wider conversation. Another part of this conversation is that it makes sense to explore the possibility that unacceptable negative economic outcomes may seriously affect a large part of society during the long run.
Concerns have been raised that if entire generations of workers face physical harm, long term joblessness, homelessness, or poverty then it makes sense to define contingency plans that are invoked if it's ever determined that AI fundamentally contributed to severe negative outcomes.
Similarly, if AI threatens serious harm to the national economy in ways that our current economic system is not prepared to handle, or poses an unacceptable threat to national security, then having plans in place to deal with these very negative scenarios makes sense. No independent economists or political analysts say these negative outcomes are impossible. This should persuade us to start looking at additional and responsible policymaking now. If these outcomes never occur then no contingencies are triggered, no innovation is inhibited, and progress is allowed to continue and is encouraged.
This is similar to events that happened during the COVID outbreak when government stimulus money was sent to residents because politicians recognized that there was a need to avoid economic collapse, and there were defined contingent policies already in place that the government could fall back on.
Defining these temporary or possibly longer term plans that use similar existing contingencies that are already in place, allows solutions to be available to prevent negative outcomes like economic collapse when triggered. Plans, such as similar insurance options and safety nets, are far more effective if put in place early, before serious negative events are triggered. This type of proactive reaction is intended to preserve better outcomes for workers and communities during clearly identified and seriously negative events.
Studying historical accounts, like the Industrial Revolution period, and other eras that incorporated technological advances, shows that corporations are not motivated to develop contingencies before serious negative problems occur. There needs to be independent third party sources proposing possible solutions. For example, the FAA is an independent third party organization that examines new airplane models before passengers are allowed to board and begin flying. The independent third party organizations exist because corporate strategies are driven by profits that benefit shareholders. This alternative is intended to prevent serious harm to families and communities during the long run.
Finally, there are several serious possible problems facing all of us today. Is it possible to rank the risks of AI when compared to other serious problems? Some of the most serious risks involve dystopian results like pandemic outbreaks. Monitoring the likelihood that any of these results will spread too far is worthwhile. People are already convinced they're worthwhile actions. Studying the possibilities does not threaten anyone. No one believes it's a bad idea to study possible negative outcomes that might occur when rocket ships are sent to Mars, even though a catastrophic failure would not seriously harm the national economy, nor create a serious national security issue. The biggest concern I have is that it appears to be quite difficult to convince the government that these studies involving AI are needed, and that well structured contingencies will do far more good than bad.
Take a moment and compare both positive and negative AI outcomes to what I believe is the most pressing issue we face today: the concentration of wealth into the hands of just a few people. No matter what the answers turn out to be to questions about AI making the lives and jobs of workers better or worse in the long run, those answers will not seriously affect wealth concentration. The concentration of wealth will continue if AI turns out to improve lives, and it will continue if it does not. Evidence shows that the concentration of wealth is already contributing to the increase in homelessness, the increase in poverty, and the decrease in the number of families achieving middle-class lifestyles. This increasing degradation of lifestyles has been happening for several decades and is continuing. It's even possible to envision AI contributing to wealth concentration accelerating at a faster rate. Perhaps policies are needed that will be triggered if wealth concentration continues to lead us in very negative directions. Everyone needs to think about what policies are needed if there are serious triggers and it's clear policies should be activated to prevent unwanted societal collapses. In conclusion, there are three kinds of people, those that believe wealth concentration leads to poverty, lack of food, and healthcare insecurity, and those that don't, and those that say it's worth studying.
See No Evil, Hear No Evil, Speak No Evil (AI)
I think both AI and the concentration of wealth are worth studying and I believe most people, when they look closely at the evidence, will come to the same conclusion. Speak up!