In his September keynote address at an event in Mexico City, Microsoft (MSFT) CEO Satya Nadella explained why the company had made another $1.3 billion investment in the country’s artificial intelligence programs: “We are entering a new era of A.I.,” Nadella said. “With the promise to create inclusive economic growth and opportunity across every role, industry, and country.” The dawn of this new A.I. is transformative. Our tools are powerful enough that ideas that once took years of repetitive, mind-numbing development can now be accomplished in a few weeks. If you think of a way to make our lives easier, richer or more rewarding, it can be implemented (even with limited resources). But to take advantage of this moment, we must get a few things straight.
Decades of science fiction have primed us to fear a technological revolution. It’s normal to worry about your role being made obsolete by a line of code, especially as we see A.I. do things that seem almost magical. But if we learned anything from the introduction of manufacturing robotics, there won’t be some wave of mass unemployment. Whenever we build a new technology, it breeds new opportunities and industries, often creating more jobs than it eliminates.
It’s called “creative destruction,” and there are already signs that A.I. will be no different than the advancements before it. The World Economic Forum’s whitepaper “Jobs of Tomorrow: Large Language Models and Jobs” identified several areas that show strong growth potential—namely, ones that require abstract reasoning and critical thinking skills—and others that are relatively unaffected.
While the WEF’s “Future of Jobs Report 2023” projects a decline in data entry jobs, it also expects increases in electrotechnology engineers and sustainability specialists—industries with a significant opportunity for A.I. augmentation. We will lose some repetitive, low-complexity jobs to A.I. agents. But those agents also need humans to build, guide and refine them. Those agents can help support the exciting imagination and innovation that moves our society forward. Want to make art? Make it faster. Want to build electronics? Perfect your design with no trial and error. Want to protect the environment? Let A.I. predict forest fires and better direct scarce resources.
We will lose some repetitive, low-complexity jobs to A.I. agents. But those agents also need humans to build, guide and refine them. Those agents can help support the exciting imagination and innovation that moves our society forward. Want to make art? Make it faster. Want to build electronics? Perfect your design with no trial and error. Want to protect the environment? Let A.I. predict forest fires and better direct scarce resources.
There are different ways that A.I. can help, but make no mistake—it does help. In their dramatically titled paper “Navigating the Jagged Technological Frontier,” researchers from Harvard Business School wanted to see the effect of A.I. on an industry often considered at risk of obsolescence from this new technology. The role of ‘consultant’ is, by nature, hard to pin down. There are many different flavors of consultancy, from McKinsey to Infosys. But in general, someone takes in a large amount of information, breaks it down into the most important bits and then makes suggestions based on historical data. That’s precisely what A.I. does well. So, when Harvard researchers gave A.I. tools to consultants, it produced some understandable results. Those with access to A.I. produced 40 percent higher quality results than the control group (those without A.I.) and consistently outperformed their previous results.
Interestingly, while these were study-wide, two distinct groups emerged. They were dubbed ‘Centaurs,’ who divided their tasks between themselves and the A.I. (like the half-horse-half-human mythological creature), or ‘Cyborgs,’ who integrated the A.I. into their entire workflow. It’s not clear which method is more effective—something fascinating to me—but it is clear that A.I. improves in either case. Like any study, this one comes with a caveat. When the researchers gave the A.I.-augmented consultants tasks that required information or data that the model didn’t have, performance dropped. The A.I. gave incorrect help, and it wasn’t scrutinized well enough.
Since it is almost impossible to feed an A.I. model all human knowledge without making it gigantic, costly and unwieldy, there is great potential for “domain-specific generative A.I. models” to become a core focus. These specialized models are optimized for specific fields of expertise and reduce the need for advanced prompt engineering, as seen with general-purpose models. By training the models on more targeted datasets, the risk of hallucinations is predicted to decrease significantly. Given this context, further technological development in this area is expected to continue, and the current challenges will likely be tackled over time. These results highlight the continued importance of human oversight in the validation process and reveal some of the current limitations of A.I.
The question then becomes whether our focus should be on simply training employees to use A.I. to improve their results or developing new processes, products and innovations. Realistically, it’s not an either-or scenario. Both things will naturally happen, but the latter will truly prepare the next generation of workers. Take the partnership between Fujitsu and MoBagel as an example. Combining Fujitsu’s robust infrastructure with MoBagel’s innovative platform, it has rapidly developed and deployed A.I. enterprise products. There are brilliant young minds who know how to leverage these kinds of technological synergies to create groundbreaking solutions. But to fully reach that potential, our educational and professional training systems must evolve.
Our current education system needs to better prepare students for the A.I.-driven future. The classroom needs an overhaul that emphasizes critical thinking, creativity and collaboration over rote memorization. A.I. literacy should be integrated into the core of our educational institutes, alongside ethics, humanities and social sciences. The goal should be to ensure that the next generation of innovators is technically proficient and socially responsible.
A survey from IBM showed that 40 percent of the current workforce will need to reskill due to A.I. implementation in the next three years, but it’s more than just training them how to use it. Workers should be encouraged to learn more than the basics. We need to instill a culture where continuous learning and adaptation are the norm, not the exception. As business leaders, we must invest in programs and opportunities for employees to upskill, not just reskill. Governments, too, have a role to play through public policy and funding initiatives supporting formal education and informal learning opportunities. There’s no denying that the landscape of work is about to change. But that doesn’t have to be frightening. A.I. allows us to do things we previously thought impossible with fewer resources than we ever imagined. Those resources can be rerouted to support innovative new products, ideas and operational methods. If we prepare our employees properly with a new educational paradigm—both in school and the workplace—focused on capturing the limitless potential of A.I., it won’t be a revolution; it will be a revelation.
