Anticipating the Impacts of AI on Jobs, Culture, and Society
|
A drag holds a human child in Fantastic Planet (1973) |
Artificial Intelligence (AI) has finally arrived.
With the release of ChatGPT3 at the end of 2022, a form of so-called artificial intelligence has become part of the lives of millions of people around the world.
What was once an academic curiosity has now become a dominant mega-topic. The official birth of the AI field is considered to be the legendary Dartmouth Conference in 1956, where John McCarthy coined the term "Artificial Intelligence." Alongside Marvin Minsky, Claude Shannon, and Nathan Rochester, McCarthy organized the conference, thus establishing himself as one of the grandfathers of a development that today stirs the minds of many.
This success was by no means assured. From the Dartmouth Conference to ChatGPT, it was a story of ups and downs, where peaks of minor successes were immediately followed by troughs, the so-called AI winters. After exaggerated expectations that simply could not be met, markets reacted with disappointment, and financing for this esoteric branch of research was cut—leading to a refocus on tangible, achievable tasks.
Now, however, AI has entered the mainstream. It is ubiquitous, dominating investment decisions, celebrated by some as a harbinger of a future of abundance and feared by others as the last invention of humanity before its unstoppable demise.Once again, we are dealing with exaggerated expectations. Yet, the world of AI experts agrees: there will not be a third AI winter. The genie is out of the bottle and cannot be stopped.
But can we be so sure of this? Could the enormous data hunger for training AI models and/or the horrendous energy demand associated with their use, along with major accidents, simply exaggerated expectations, or other, yet unforeseen events, not again act as a brake and lead to disillusionment?
I consider this possible. But here we are concerned with something else.
"We have reached the threshold of a development that will sustainably change the world in which we work and live."
This statement has been heard frequently in recent weeks and months in one form or another. If we follow this assumption for a moment, with all healthy scepticism towards forecasts, what subsequent effects can we then expect for jobs, culture, and society?
A look at history teaches us that human development proceeds in waves. By no means does it follow a straightforward trend towards more, however defined, "progress." Let us assume for simplicity's sake that the development of the past few years continues in the same direction and speed as before, or even faster. Then AI will indeed sustainably change our lives.
But how?
This is something we need to think about.
This raises a multitude of questions that are not easy, perhaps even impossible, to answer, but whose answers depend on our fate as humanity in its entirety.
Questions such as...
- Are we approaching a time of abundance, as enthusiastic futurists like Peter Diamandis have been predicting for decades? Or do we first lose our jobs, then our purpose in life?
- Will human and machine share work amicably, so that each party does what it is best suited for? Or will AI colleagues soon do everything better than us, and human intervention would only interfere?
- Will a new era of machines dawn, a level higher, more powerful, and more productive than today? Or will machines develop into our new colleagues? Will they, a creation in our image, have consciousness, their own opinions—a will of their own?
- Suppose the development occurs eruptively and quickly in a few epicenters of our planet. Will we find the right solutions for societies in individual countries of this earth and between states for adequate participation in the fruits of this development?
- Will we have enough time to react to these changes? Do we perhaps even have to anticipate them as alternative scenarios? Or will we, as so often in the past, only notice the societal consequences when they can no longer be denied, and a revolutionary wave demands economic, societal, and thus political changes with disruptive force?
Much about this AI-driven innovation may be new, specific, and quite different than before. Yet, if we want to look into the future, a glance in the rear-view mirror often helps, at past economic waves, often referred to as Kondratiev cycles after their namesake, the Soviet economist Nikolai Kondratiev.
What advice do we receive from a glance at history when we ask it three simple questions?
- What technological innovations in the history of humanity had societal and political effects?
- What were these effects?
- How long did it take until these effects became noticeable?
We should be aware that the impact of technical innovations is often overestimated in the short term but tends to be underestimated in the long term.
What does the retrospect teach us?
Throughout human history, there have been numerous technical innovations. The economic effects of the aforementioned 5 Kondratiev cycles are comparatively well studied. There is significantly less consensus about the societal and political consequences. There had also been innovations before the beginning of industrialization that had significant societal and political effects.
Again, three of them I will consider here as examples. They include...
- The printing press
- The steam engine
- The internet
The Printing Press and the Explosive Power of Knowledge for All
|
Johannes Gutenberg |
Developed by Johannes Gutenberg in the 15th century, it created the conditions for previously insider knowledge to find widespread dissemination in large quantities.The effects of the innovation "printing press" were noticeable within decades.
Favoured by a political environment in which the secular power of traditional authorities was weakened in the aftermath of previous, catastrophic plague epidemics, we saw as immediate or indirect consequences far-reaching intellectual and political movements, such as...
- Renaissance,
- Humanism,
- Reformation,
- Enlightenment
This "emergence of man from his self-imposed immaturity," as Immanuel Kant would later call it, had undeniably positive effects, but also side effects that would plunge Central Europe into a catastrophe, the consequences of which were felt for a long time.
The newfound freedom of thought eventually shook divinely perceived social orders. Counterforces were not long in coming. In the fateful 30 years between the Defenestration of Prague and the Peace of Westphalia, Central Europe was largely devastated. This Thirty Years' War reduced the population of Central Europe by about half.
The Steam Engine and the Long Shadow of the Industrial Revolution
|
James Watt |
Linked with the name James Watt, the steam engine was developed in the 18th century. According to conventional doctrine, it thus initiated the industrial revolution, the 1st Kondratiev cycle.It changed working conditions and economic structures sustainably, leading to a wave of urbanization. Per capita incomes in the USA, for example, doubled during the 19th century. Life expectancy in Great Britain rose from 40 to 48, the literacy rate from 75% to 97%.
These societal changes became apparent within a few generations.
But this time too, the again undeniably positive effects had their clear downsides. The resulting population growth increased the supply of cheap labor, which was consistently exploited. Social inequalities, exploitation, and the alienation of the working class were the result.
In 1867, Karl Marx published Vol. I of his main work "Das Kapital." He thereby created the theoretical foundations for socialism and communism. The causes, however, lay in the societal tensions enabled by an innovation and the political forces thereby unleashed. They caused revolutions, counter-movements, and ideological conflicts that are still palpable today.
The Internet − the Infrastructure for Communication and Interaction
|
Tim Berners-Lee |
Since the 1960s, developed out of a military need and initially used civically in the academic sphere, the internet has triggered a 5th economic cycle according to Kondratiev.It has profoundly changed the way we communicate, do business, and organize political movements.
The effects became apparent relatively quickly, in the late 1990s and early 2000s.
As of today, it offers almost every citizen access to the world's knowledge, to communication with individuals, interest groups, and companies. More than in the developed world, it has meanwhile enabled several billion people in the still developing world, such as in India or China, to participate in secured financial transactions and interaction with authorities and companies.
The downside, however, is that it is also used for propaganda, fake news, conspiracy theories, and hate tirades. Although these threats to communities, whose basis is reliable information, are as old as humanity, the now available amount of information is new. Our lack of familiarity with conscientious source verification proves to be problematic here. Increasingly, the abundance of information on offer overwhelms the individual citizen.
Thus, the ubiquity of the internet also tempts surveillance and manipulation by friend and foe. What once began as the great and for official bodies uncontrollable free space has meanwhile developed into the most monitored communication medium in the world. States and companies that operate a significant portion of the internet infrastructure have thereby gained enormous power over states, companies, and citizens who only use this infrastructure. As expected, counter-movements are forming here as well. Their influence, however, is not yet foreseeable. The race is still open. We find ourselves in the midst of a breath-taking development.
These innovations show how technology can change social and political structures, often faster than societies can prepare (for them).
What does this mean for the effects of AI?
Since we have been in the age of artificial intelligence for a few years now, I have, out of courtesy, first also asked AI about its own concerns. So what does ChatGPT 4.0 think about this? − Let's give the colleague (unprompted) the floor.
The question was:
If we transfer these experiences to artificial intelligence.
What can we then expect?
The somewhat sobering answer is given below along with my somewhat flippant comments:
When transferring these historical experiences to artificial intelligence (AI), we can expect significant societal and political changes.
⇨ Oh, really?
AI has the potential to reshape labour markets through automation, increase efficiency in many sectors, and fundamentally change the way we process data and make decisions.
⇨ We didn't need to ask ChatGPT for that
The effects could include both positive aspects such as improved quality of life and healthcare as well as challenges such as job losses and data protection issues.
⇨ I would never have thought of that
These changes could become visible relatively quickly, as the technology develops and spreads rapidly.
⇨ That's true. So, which AI are we talking about?
These answers could have been spontaneously provided by a mediocre career politician. Perhaps they could also have been delivered by t some phrase generator. I guess I'll have to, according to Kant again, dare to use my own mind.
ANI, AGI, ASI − AI is not all the same
ANI - AGI − ASI, these three acronyms are commonly used to describe the stages of an expected development of artificial intelligence. This is - counterforces unconsidered - the expected course.
The three stages listed are briefly characterized as follows.
ANI − Artificial Narrow Intelligence
- Models trained for specific tasks, which can already be far superior to current human capabilities in those tasks.
- This is the position where development has arrived here and now. Its influence is already palpable in approaches today but will intensify significantly.
- ANI will sustainably transform the economy and society. We will feel effects, perceived as positive as well as negative, that are comparable to those of the three selected innovations, or may even exceed them.
Even at this stage, a technology impact assessment is an almost impossible task. However, not addressing it would be a culpable omission.
AGI − Artificial General Intelligence
- This is the state popularly referred to as "Singularity." Then machines will reach human level. Since there is no consensus on exactly how to define that, I add "in all intellectual and emotional dimensions".
- Thus, they should have left the level of a (soulless) machine and mutated into a new life form, an act of creation in which we create a being in our image, if that will ever be possible. These beings will have consciousness, must receive rights & will (unfortunately) be similar to us.
- They will be as smart as an average human, but also as stupid. They can then be our partners, assistants, or colleagues on the job − but also autonomous warriors. And of course, they will also take their own further development into their own hands, be able to reproduce independently.
ASI − Artificial Super Intelligence
- An ASI, probably a "being," is by definition far above the level of human intelligence. Since we have no model for such beings, I'd rather skip a further description here.
- It is also completely unclear whether this state is fundamentally achievable. Whether it is even desirable is another philosophical question. Some recognized AI experts and successful book authors think that the open-ended ASI universe can be reached very quickly once we have reached the AGI level. But in the graph above, the innocent term "unknown breakthroughs" stands for this. So, fundamental technical and perhaps even epistemological advances are still to be achieved, whose nature we cannot even guess, before the AGI stage can be reached.
- For decades, prominent AI developers and researchers have been warning that their own work will create this homunculus, which after a Eureka moment "It‘s alive, it’s alive!" will take over world domination and destroy us all or at least treat us as negligently as we treat the wildlife around us today.
- Yes, if these super beings will be socialized in the same way we humans have become over the course of our evolution, then this is a valid option, albeit not a very pleasant one. This horror scenario is of course pure speculation. Perhaps a real superintelligence would also have completely different priorities. Maybe it would even be the salvation from ourselves.
ANI − Tense in the Here and Now
Jobs − Will we all become unemployed?
Let's first hear the voices of others. What do our recognized AI experts have to say on this topic?
| Raj Mukherjee |
|
"We've seen it all before.
Every technological revolution has led to the loss of jobs.
But new jobs will be created, and many new jobs will be created in the course of this AI revolution.." |
Raj Mukherjee thus represents the prevailing opinion. According to this rosy future picture, humans and robots will divide the work according to their capabilities: we humans will keep interesting, fulfilling tasks for ourselves. Everything annoyingly boring, repetitive, mechanical remains with the bots.
However, this utopia is spontaneously opposed by three caveats. Because ...
- Even today, AI has a firm place in the creative business.
- Just for simple tasks, the human is simply still simply the cheaper option.
- For some "bullshit jobs" one may simply not want to admit that machines can do that too. Here, a human must still "stick his neck out".
Maybe this time, however, the course will take a different direction.
According to an analysis commissioned by the IMF, 40% of jobs worldwide will be "affected" by AI. In developed economies, it is even supposed to be around 60%. Classic industrial jobs will hardly change thereafter through the use of AI.
Academics and office workers in manufacturing, however, will be supported by AI. As a result, fewer of them will be needed. Other corporate tasks can be completely taken over by AI.
Well-founded and painstakingly acquired knowledge through years of study will no longer be a job guarantee. As long as the threshold to AGI is not crossed, human labour will certainly still be needed in many places in our economy and administration. There will be fewer of them than before. On the other hand, some new positions will be created.
But let's assume that AGI will actually be reached at some point, then it will still take a while. But at some point, we will all lose our jobs.
Inequality - It has the potential to tear societies apart
Let's first hear the voices of others. What do our recognized AI experts have to say on this topic?
| Sam Altman |
|
"I believe that AI will be the greatest force for economic development.
We will see many more people become rich than ever before." |
What will become of the rest, I involuntarily ask myself, the "left behinds," how will their live be?
Without far-reaching interventions in our way of life and economic model, the application of artificial intelligence will probably only further intensify the already prevailing trend towards economic inequality. This inequality usually grows in politically and economically calm times.
Without intervention it can take on extreme dimensions and can only be reduced by catastrophically incisive events such as epidemics, wars, or revolutions. In doing so, economic inequality grows within national societies as well as among nations, if they do not fight back with means that are considered unfair by convinced market economists.
Extreme economic inequality can lead to social unrest and economic inefficiency. It has the potential to tear societies apart. As with previous innovations, there will have to be political responses to such technology-induced upheavals.
Unlike in the past, however, we will not have much time to address the challenge. We have not even generally recognized it as a danger yet.
Changes in Culture & Society
"The future is here − it is just not equally distributed" is often quoted from science fiction author William Gibson. That means, if we want to know what it will look like (soon) in our area, we just have to look around. In a few, select avant-garde corners and crannies of our society, we will find material to view for a glimpse into our own future.
Yann LeCun, AI expert at Meta sees AI as the future of communication: "We will constantly talk to these AI assistants. Our entire digital communication will be mediated by AI systems." Will the already difficult communication between people then fall by the wayside? Well, it will inevitably happen less frequently in this scenario. Hopefully, we won't get out of practice.
Scott Adams, the well-known author of the popular Dilbert cartoons, even sees the gender ratio disturbed. From his (male) perspective, men will increasingly turn to low-maintenance digital girlfriends in the future: “The girlfriend Singularity is here. Human women had a good run.“
This may still seem quite far-fetched to us. Nevertheless, our interaction will change. Rather questionable offers in this context already exist. Apparently, we are in for a journey into the unknown.
Another question we will have to deal with is how we will deal with the loss of a profession that fulfils us, makes us proud, from which we derive our self-esteem. Is a life without a mission meaningless? Will decreasing self-esteem due to job loss become a mass phenomenon?
The much-discussed unconditional basic income may become necessary. As long as our social status is job-centered, and just as our self-esteem is defined by it, however, it will not solve the problem of job loss.
In our lives, other values will have to gain importance. Maybe we should look to role models in other, foreign cultures or on the fringes of our own societies, for whom work has never been the centre of life and who have nevertheless been able to conduct a good life.
We will have to redefine the meaning of life.
Outlook on AGI & ASI - What Developments Are Expected?
Here we enter unknown territory. This is the realm of pure speculation. Some well-known AI experts have already undertaken this for me. Let's let four of them have their say.
| Elon Musk |
|
"For the first time, we will have something that is smarter than the smartest human. It's hard to say exactly when that will be, but there will come a point when no work will be necessary anymore." |
| Yuval Noah Harari |
|
"We may be talking about the end of human history - the end of the period dominated by humans.."
|
| Paul Christiano |
|
"I think there is a 10- to 20-percent chance that AI will take over, with many, most people dying."
|
| Geoffrey Hinton |
|
“Humanity is just a temporary phase of evolutionary intelligence.“
|
Will an Artificial Super Intelligence (ASI) deal with humanity as we deal with the wildlife today?
Is it even responsible for the "Great Filter" resulting from the Fermi paradox, which has so far prevented aliens from contacting us (because they have all been turned off by their own AI beforehand).
So exciting philosophical questions arise. Personally, I tend to agree with Kai-Fu Lee, book author, former Google China boss & founder of the start-up 01.AI: "The Singularity is still not our problem".
But we have to deal with the immediate consequences of a broad application of artificial intelligence, with wealth concentrations, possibly mass unemployment, and the loss of a natural self-understanding that was based on one's own creation of values.
If we have indeed reached the threshold of a development that will sustainably change the world in which we work and live, it is now time to think about the consequences for jobs, culture, and society.
As in the wake of the great innovations of the past, there will have to be political responses to the new economic constellations.
Unlike in the past, however, we will not have much time for an appropriate response. And there will be political responses, whether we act now or not. It would therefore be good to think about the "how" now, to at least have the chance of influencing the course of history ahead of us.