AI is often regarded as the greatest invention since electricity. A bold claim, but one not unfounded. The potential of AI, generative or physical, is only limited by human imagination.
It is a technology that can define us as a civilization. For it gives us the power, however minute it may be, over the thing we covet the most—time.
The common consensus among workers is that in the not-so-distant future, AI will gain the capability to automate all existing jobs, leading to the obliteration of the global workforce.
While this is a plausible forecast, grounded in maths and AI capabilities. It is simply a case of misplaced fear.
THE PAST.
In 1985, Kary Mullis, a biochemist at Cetus Corporation, invented the Polymerase Chain Reaction(PCR). A method used to make billions of copies of a DNA sample rapidly. This seemingly simple procedure has been described as the “electricity of modern biology and medicine”.
A definition well deserved considering the overreaching effects of PCR in medicine, biology, DNA study, and forensic science.
However, all this would not have been possible if Mullis had not been affected by the cold, unfeeling nature of automation.
You see, before Mullis's invention of PCR, if you wanted to study a DNA sample, you had to break down the DNA into fragments using what were called restriction enzymes.
Through a method called Gel Electrophoresis, the stripped DNA fragments are segregated by size, after which a Probe(a separate DNA sequence complementary to the desired DNA fragment) is introduced that will bind to any matching DNA fragment(s) with a complementary sequence.
Finally, the DNA sample is washed to remove any unbound DNA fragments, thereby isolating the intended DNA fragment.
This procedure is called Southern Blotting, a tiresome, slow process that was grossly ineffective.
At Cetus, Mullis was in charge of manually synthesizing the required DNA Probe—a crucial step in the Southern Blotting procedure.
It was not an enjoyable job; it was repetitive and slow, and perhaps as a result, he was known to have quite the temper, often getting into fights with his colleagues.
He would continue this repetitive task daily until the invention of an automation tool for synthesizing probes. It was called the synthesizer—a creative name.
All of a sudden, Mullis had nothing to do; he would spend most of his time tinkering with these DNA samples or taking the weekends to relax in his cabin in Mendocino County, California.
On one of those occasions, driving home, he let his mind wander on the current method of DNA tests. All of a sudden, he had an eureka moment. He thought of a solution that could solve all current constraints of DNA study. This solution, he would name the Polymerase Chain Reaction.
In his own words, he knew right then and there, that he would go on to win the Nobel Prize for this discovery. He did.
It is easy to attribute this breakthrough to his genius; some might even say it was a coincidence. Mullis had even stated in several interviews that the source of this discovery was his addiction to LSD(addictive drugs).
However, there is one underlying factor everyone disregards. One equally important factor as LSDs, that led to the discovery of this miracle solution, was automation.
It was automation that made Mullis idle; it was automation that allowed him the privilege to take weekends off when others were confined to their labs.
Automation took his job, but gave him the freedom and time to be creative, to reinvent, to change the status quo. This is the real power of technology.
If we flip through the pages of time, we find out that all technologies with a global impact saw a period of mind-blowing inventions, led by creative minds, soon after
As far back as the 18th century, during the Artisan Era, goods were strictly produced by artisans, experts with keen judgement and years of experience in their field. Firearms were produced by blacksmiths, clothing by tailors, shoes by shoemakers. It was simply impossible to live without artisans, and their worth was indispensable. This, however, could not satisfy the needs of a growing global economy. Goods were of high quality, but there was a shortage of them. This constraint required a change.
Then came the Industrial Era. What took artisans years to master was broken down into small sequential steps that could be imitated by sophisticated machines alongside human input. This led to an unprecedented boon in productivity, the likes of which had never been seen before.
There was once again a looming constraint. Factories and companies grew in productivity to the point that there was no way to effectively manage production, labor, and logistics on a global scale, while staying up to date with competitors and new inventions. There was a need for a more effective way of communication. This led to the creation of tools like telegraphs, typewriters, and tabulating machines; the earliest form of computers, the internet, and communication devices, leading to the Information Era.
In all these Eras, new tasks/Jobs and innovations were created as a result of these technologies.
Technology Life Cycle: https://generativeai.pub/ai-doesnt-steal-jobs-it-creates-new-ones-e1453ca3b62a
Without machines, the global applications of electricity would be impossible. Without electricity, the internet would not exist. Without machines and electricity, computers would not exist, and without the invention of all three, AI would be impossible to achieve.
How, then, does AI affect our lives now and in the future?
THE PRESENT.
We often define a job as if it were a single task assigned to an individual. That is entirely incorrect.
A job, in simple terms, is a bundle of tasks. These tasks are usually interconnected and are carried out sequentially or simultaneously.
Among these tasks, there are two major categories: Easy and Hard tasks(Acemoglo et al.
Unlike the names, these tasks may or may not be easy to learn, but that is irrelevant. These tasks are rather categorized based on two factors: The results and the process.
For Easy tasks, AI can learn to replicate them because the results are controllable, and the process gives you the expected result. Examples include data entry, account query, coding, etc.
Hard tasks, on the other hand, are defined with uncertain results that rely on creativity, and results are heavily affected by context. These tasks require abstract reasoning in complex environments that could affect the outcome of the task. A chef preparing a meal, a doctor performing surgery, and a manager making an impromptu decision based on findings are some examples.
In the present, Generative AI like ChatGPT and Claude can perform Easy tasks efficiently without human input, and oftentimes the quality of such tasks is much better than that of humans.
This result is quite difficult to achieve for Hard tasks, due to the contextual nature of such tasks. There is, however, a caveat to this reality, which is that this technology is constantly evolving, and results will continue to improve as time goes on.
In understanding this, we can define the current state of our interaction with AI technology.
Fundamentally, we implement AI in two ways:
The ability to automate tasks is not a novel experience, as we have done since the Industrial Era. Before the advent of Generative(multimodal) AI, automation already existed. These automations were strictly rule-based and followed predetermined instructions to perform tasks. They had no element of adaptability and could not learn or reference.
While they were not intelligent, they performed efficiently as long as no complex or contextual layer was involved in the task, and they are still widely used to date. Such examples include assembly lines, traffic light systems.
AI enhances these models to be able to infer, learn, and adapt based on available data. We see companies invest in AI-powered robots to carry out factory work, navigation apps like Google Maps, and Smart home assistants like Alexa, among others.
Self-driving cars, virtual assistants, document processing, email filtering, robot guidance/coordination, bug testing, network test detection, writing, and content moderation on social platforms are just among the diversified uses of AI automation.
Filtering through the examples, it is obvious that not all examples are Easy Tasks; examples such as self-driving cars can be quite complex, yet AI automobiles like Tesla perform just as well as expert human drivers.
Everywhere you look, there is some element of automation, at work, at home, at school; there is no part left untouched by automation, and AI built upon this penetrating foundation is the logical path.
In this application, AI does not replace human tasks, but rather complements or enhances them, improving the quality and reducing the time taken to complete them.
Content generation, vibe coding, academic research, and graphical designs are among the many examples where AI provides possible paths, creates drafts, and proposes ideas, while humans finalize the output.
A crucial characteristic is that this application has the biggest impact on low-end or entry-level workers, allowing them to bridge the gap in skill and experience against senior workers.
In a research carried out by MIT graduates( Noy, Shakked and Whitney Zhang), a group of workers were randomly selected from a range of white-collar occupations and presented with simple writing tasks(familiar tasks to their jobs) that would take 30 minutes to complete.
The participants were separated into two groups: the control group, which was restricted from using AI tools, and the treatment group that was encouraged to use AI tools like ChatGPT 3.5 to aid their writing. The results showed that the ChatGPT group had a 40% increase in completion rate and an 18% improvement in quality scores. In both cases, the majority of the gains came from workers who performed poorly in this task before the experiment.
Both applications of AI are a result of the rate of adoption of this technology, and will not be the only application of AI as adoption increases and more of the world gets exposure to this viral technology, and it may not be so far away.
According to a recent research carried out by the International Labour Organization(ILO), one in four jobs has exposure to AI, and one in three jobs in high-income countries are exposed to AI.
The most fascinating thing, perhaps, is the knowledge that Generative AI, the main protagonist of this exposure, has only existed for less than three years, counting since the time ChatGPT first launched in November 2022.
Since then, we have witnessed an explosive rate of adoption that has never occurred with past technologies. This rate of adoption supersedes that of computers or the internet.
This is partly due to the already existing layers of technology AI is built upon: the internet, computers, and electricity. This means it faces none of the limitations that plagued its predecessors, but rather absorbs their strength; the computer brain and the sea of information that is the internet.
Yet, this adoption doesn’t seem to be slowing down, with a net increase in the number of users globally compared with last year’s result. AI is quite literally the future, and rather than fight against it, run from it, we should seek to understand it, explore its use cases, and adapt to it.
Perhaps, we may learn a thing or two from AI.
”History tells us that the road to technological adoption is bumpy”
Technology is the foundation for increasing the standard of living. It can also be the butcher’s blade that disrupts the economy by displacing jobs, favoring the rich, and widening the gap in education and demography.
AI can also be the foundation for our civilization’s progress, or the blade that strikes an uncrossable chasm between the rich and the poor, the privileged and the less privileged, the noble and the common.
At this rate of technological adoption, the future is uncertain. Yet, we are the ones who will have to shape this future. Therefore, we should be concerned and conscious about how this technology will impact our lives and that of those yet to be born.
It is easy to leave it up to fate, to go with the flow, but history tells us that may not be the best option.
We went with the flow during the Industrial Revolution, leading to power concentrated in a few hands, loss of jobs, inhuman exploitation of workers, and social inequality.
We did this in the Information Age; the implication? Media became a controlled and regulated tool by the powers that be, we saw the exploitation of human privacy, we witnessed a digital divide, and abuse of user data, among many others.
Now, we are at the same crossroads: to be active or to be passive, what choice do we make?
We may not be certain about what the future holds for this budding technology. Still, we can, by reference, project possible scenarios by looking into the path we have taken in adopting human technologies. Here are the possible paths we can take:
Imagine a future where the majority of companies and organizations adopt AI to reduce production costs. In this future, AI does the majority of tasks in a company or organization, aside from key decision-making tasks.
What this means for workers is that companies will be offloading their workforce to adopt the less expensive, more productive alternative—artificial workers.
While this may sound pessimistic, it is a rather sound conclusion. Companies, in essence, seek always to increase productivity and reduce costs. AI offers both. It is far more efficient at handling tasks and far less costly. It is not pessimistic, but the reality we live in.
This path will mostly depend on government policies and companies’ goal towards AI adoption.
It is also important to note that this massive layoff may not happen all at once, but will be proportional to the advancement in AI technology. Quite frankly, AI cannot fully automate the entirety of existing jobs, but this will not be the case forever.
In this future, we take the more optimistic approach. Companies provide specialized training to their workers by educating them on the applications of AI in their workplace, thereby increasing the efficiency of their work.
This would be especially helpful for low-skill workers, allowing them to perform just as good as experts in their field, as seen in the MIT grad writing experiment. This, in turn, increases the value of workers, leading to higher pay and a demand for AI-savvy recruits. This may also encourage an age of AI-learning among undergraduates and young adults.
While this path sounds too optimistic, it is, nonetheless, a possibility that should not be ignored. It can be made possible if governments enact policies that encourage human-AI productivity and policies that punish the degradation of the current workforce by companies seeking the cost-effective path.
This path is estimated to be quite slow, as it takes time to cultivate a society/workforce geared towards AI-aided productivity in an age where automation is a shortcut, but the results will be worth it.
Most importantly, it requires the active participation of citizens in AI adoption. Being aware and intentional about how AI can be used and applied will be instrumental in creating an environment for support from government and companies alike.
Every period of technological progress and adoption has led to the creation of entirely new products, sectors, and markets, which in turn create new tasks for workers and jobs that pertain to these tasks.
If such is the future of AI adoption, jobs will be abundant. Looking back at the life of Kary Mullis, the inventor of PCR. Automation led to creativity, and creativity led to the breakthrough that created new branches of biological studies, and by extension, new jobs unique to those transformations.
AI has this potential, and on a much higher scale. At the moment, we can already see the formation of such new jobs. Vibe coding, as a form of programming, is a prime example. For an industry so radically affected by AI, one would expect job degradation, yet that doesn’t seem to be the case. Nowadays, programmers utilize AI in the same way they utilize VS Code, as a tool to increase their work efficiency greatly. This combination is called vibe coding, and it's been used on all levels of programming.
This example will not be the last, as AI continues to develop, and more users find creative ways to apply it to existing tasks, we will see a transformation of jobs that would make workers much more efficient and productive.
This path is beneficial to companies and individuals alike, but will only be possible with the active participation of its adopters.
It is important to note that all three paths can converge. The task-complimentary path will eventually lead to the task-creation path as individuals and companies implement AI into their work process, finding intuitive ways to improve productivity via AI.
The task creation path can lead to the full automation path. With the development of AI technology, there may come a time when AI achieves human intelligence of all faculties, popularly known as Artificial General Intelligence(AGI). At this level, AI will be capable of performing any tasks with the same level of intelligence as humans.
This means it would be able to learn Hard tasks, reason, and grow through a constant feedback loop of trials. In this future, AI would outperform every expert in any field as long as it is given enough time to learn.
Therefore, whatever path we take, jobs will be displaced, transformed, and created, as with every other technological adoption.
That is why it is imperative that we, as humans, take an active role in shaping this technology. We can do this by educating ourselves and others on the applications of AI, becoming an advocate for the enactment of policies and regulations that create a fair and accountable adoption, speaking up on concerns, and offering feedback to developers of AI tools and policymakers.
We now stand at the crossroads. What choice will we make?
”We should ask not what AI will do to us, but what we want it to do for us.” (David Author. NBER)