Unpacking the AI paradigm shift

Is AI similar to or different from other technological revolutions? Below, I dive deeper into AI's impact on productivity and employment, drawing lessons from history.

One of the things that I have found myself absorbed with and reading intently about is AI: How it is similar to and different from other industrial revolutions and how it will change the future of work. My initial view was that the AI revolution, similar to other technological revolutions, would make us more productive and despite disrupting some jobs, would be a net job creator. The thinking was rooted in the study of the impact of industrial developments in the past. However, there are compelling reasons that support the belief that this time is different. Let’s look at all this in more detail below.

Looking back

One of the most significant developments of the Industrial Revolution (1760s - 1840s) was the steam engine. While this invention disrupted some traditional jobs: artisans and skilled tradespeople were displaced by mechanization and factory work, it also boosted production and productivity as machines outpaced human output across the board and spurred innovation. According to this article by the Bank of England, GDP growth per person since 1750 has been on average 1.5% per year. At this rate, the economy would have taken 50 years to double in size whereas before 1750, it would have taken 6,000 years!

Similarly, the invention of the internal combustion engine destroyed the horse and carriage industry but created filling stations, auto repair, the possibility of living away from work, and having a larger house where land was cheaper. Interestingly, even the ski industry owes a thank you to the internal combustion engine as the latter made skiing accessible to the masses, led to investment in road development, reshaped ski resort operations, and created new jobs like snowmakers and ski instructors. This is just to say that the impact of new technologies on employment is not straightforward and hard to predict at the time of invention.

In a similar vein, a team at McKinsey studied the net impact on employment from personal computing since the 1980s and determined that while 3.5M jobs were destroyed, 19M jobs were created as a result of this technology.

These examples make many in the industry believe that AI, like other technological revolutions, will ultimately prove to be a net creator of jobs. This may be true, but it also (contentiously) assumes that this technology is intrinsically identical to others.

Could this time be different?

In 2017, Vasant Dhar, a data scientist and professor at NYU noted, while discussing the impact of AI machines on human labor, ‘Those weren’t thinking machines. This is not the same as last time.’ More recently, Ben Y. Zhao, Professor of Computer Science at UChicago, during a panel titled ‘Is AI making you obsolete’ in June 2023, remarked that the analogies to the printing press and other developments are not an adequate pair for what’s coming.

Meanwhile, global tech leaders have voiced similarly impactful opinions about AI: Sundar Pichai, the CEO of Alphabet Inc. has called AI more profound than fire or electricity, and, Elon Musk, the CEO of Tesla, at the AI Safety Summit in the UK 2023, said that there will come a point where no job is needed, and that Universal Income will be a societal leveler.

The impact of AI on jobs extends beyond the verbal predictions of AI Leaders or academics. The IMF in this paper released in Jan 2024 estimated that 40% of global employment is exposed to AI; this number is higher at 60% for advanced economies due to the higher prevalence of cognitive-task-oriented jobs. It also highlights that unlike previous waves of automation, which had the strongest effect on middle-skilled workers, AI displacement risks extend to higher-wage earners. Income and wealth inequality could increase if ‘AI complementarity’ (AI’s ability to complement human labor) is high. At the same time, sufficiently large productivity gains may accrue that more than offset the partial replacement of labor tasks by AI. Unsurprisingly, the report also highlights that college-educated workers are better prepared to move from jobs at risk of displacement to high-complementarity jobs, while older workers are more vulnerable to AI-led displacement.

What does this mean? I believe that AI is primed to make many roles less relevant, if not obsolete. It isn’t just a vertical disruption; it cuts across functions and sectors and is unfolding faster than other transformative advancements of the past. For instance, while the internet took 7 years to reach 100M users, ChatGPT achieved the same milestone within 2 months.

Cool Infographic by Visual Capitalist

Finally, this piece by Robbie Allen (CEO & Co-founder of Bionic Health) on Medium, puts an excellent structure to the problem of ascertaining whether this technology is truly different. He calls out two factors that are important determinants of how fast and how far a particular technology advances. The first factor is the barrier to entry for a single developer to create something useful; low barriers make rapid dispersion possible. The second factor is whether development on the core platform is centralized or decentralized; the latter makes it possible for a widely distributed group of people to contribute to the platform. After subjecting eight technological waves to this criteria study (desktop operating systems, web browsers, networking, social networks, mobile apps, Internet of Things, cloud computing, and AI), he deduced that only AI possessed both characteristics and had the potential to create fully distributed innovation. The article is worth a read if you’re interested.

Circling back to what this means for employment, this rapid pace of change will also be reflected in the timeline of impact. In a recent survey conducted by Resume Builder, 37% of the companies using AI revealed that the technology replaced workers in 2023. 44% of the executive respondents said they anticipate further cuts in 2024 due to AI efficiency. Of course, not all kinds of jobs are equally vulnerable to such cuts. The World Economic Forum’s Future of Jobs Survey 2023 predicts the largest job declines for data entry clerks, administrative & executive secretaries, accounting & book-keeping clerks, and cashiers, among others, while calling out AI & ML specialists, sustainability experts, information security analysts as among the fastest growing roles.

It is also reasonable to expect roles that involve significant social or emotional components such as therapists, social workers, teachers, priests, and those who are responsible for managing teams or making complex business decisions (C-level Executives) to prove resilient. Lawyers, consultants, and, financial advisors have some element of human touch and so should be safer, although the broad-based improvement in productivity from the adoption of ‘intelligent Co-Pilots’ should make more possible with less, so their fields too are not entirely immune, in my view.

Concluding thoughts

So, which is it? Will history repeat itself or will this time be different? I am leaning toward the latter. While this technology will result in productivity gains across the board, it has convincingly challenged the belief that AI impacts mainly middle and low-skill jobs. It is wide-reaching, rapidly evolving, and uniquely capable of creating a paradigm shift in the way humans and machines interact.

That said, while history may not repeat itself, it will likely rhyme. The path towards AI resilience involves embracing the technology through continuous education and up-skilling. While not all of us have to be (or want to be) AI/ML specialists to remain relevant, it is prudent to invest the time to learn how the technology works fundamentally, how it is changing our roles, and how we can use AI tools to our advantage. I also additionally support the view that generalists will benefit in the AI era, putting their agile brains and multi-disciplinary thinking to work in a rapidly changing and interconnected world. While some AI bloggers and influencers have written about this as well, the most evidence-backed and thorough analyses of this concept can be found in David Epstein’s book, ‘Range: Why Generalists Triumph in a Specialized World’. I highly recommend it!

Thank you for reading | What’s next?

This is all from my end. Thank you for reading my first-ever blog! I’ve wanted to do this for some time now but always struggled to find the time and motivation to put pen to paper. I hope you found this useful. This space is evolving rapidly, and I don’t think anyone has all the answers. My view is just one among many perspectives. Personally, the more I’ve read about this, the more I’ve realized how little I know. I’m not an expert on the matter and will often rely on the work of others to make sense of the technology, but what I am is curious, deeply interested, and, excited at the prospect of diving in and sharing my learnings. I’d love to hear what you think, where you agree or disagree and, what I may have missed!

For my next blog, I plan to cover more about the history of Artificial Intelligence with a specific focus on the question: ‘Why is this technology that was invented 70 years ago creating so much noise now?’ For this, I will cover both the hardware and software developments that have made this stunning technological advancement possible. Stay tuned!

---