As massive layoff announcements continue hitting headlines, many of us are wondering: Could I end up in the next wave? Indeed, recent numbers show that job security is not guaranteed by the big brand name of your employer, your function, or even your tenure. So what are the factors that can help secure your career? We spoke with Devavrat Shah, Andrew and Erna Viterbi Professor of Statistics and AI at MIT, who often teaches business executives about AI. We asked him for his advice on the matter.
Devavrat not only teaches graduate students at MIT but also offers courses to business executives who want to enhance their skills. This latter group of people is comprised of successful professionals with established careers. For anyone looking to future-proof their career, it would be good to ask why this group feels compelled to continue their education.
It’s a great question—a question that we wondered about ourselves before launching the first course on Machine Learning, Data Science, or Artificial Intelligence. We were extremely curious about the kind of student profiles we’d get. To our surprise, a lot of the students were already high-caliber executives, which seemed counterintuitive. So we asked them about their reasons for taking the course.
Many stated that they wanted to pivot their careers toward AI in one way or another. Some said they wanted to have a better grasp of AI and machine learning concepts to better understand where these could be applied in their organization. AI has been around for several years already, but there’s still quite a bit of uncertainty in the business community about what it means to them. Then there is a third category. These people had one of a few very specific business problems that they knew AI could help solve, but they were unsure whether their instincts were right, how to start, how much it would cost their organization, or how they would make business case within their organization.
The bottom line and key takeaways for us were:
So how exactly can they capitalize on this knowledge?
When I started teaching at MIT over 17 years ago, this domain was a prerogative of math geeks and academic researchers. We talked about hypothetical and theoretical, such as your classic roll-of-dice outcomes problems. But I was always fascinated with the possibilities of applying these concepts to real-world use cases.
Fast-forward to nowadays, statistics (or AI as it is currently referred to) has become ubiquitous and irreplaceable. Think about this: your phone can talk to you, your vacuum is spatially aware, and your favorite content on streaming platforms is curated just for you. These are just a few everyday examples of AI/machine learning algorithms taking customer experience to the next level.
However, this is just the tip of the iceberg of AI applications. Asking where it can be applied is the same as asking where you can apply electricity. The answer is everywhere! Just as electricity is used to power pretty much any aspect of our life—from making your morning coffee to watching TV—data analysis is used to power decision-making. It will be used exponentially more with the rise of AI, or simply put, more robust statistical models.
It’s true that big data gave new breath to the field of AI. But here’s what’s interesting—you don’t even need big data anymore to use it. There are some algorithms like multivariate time series forecasting or models that are already pre-trained on some open-source data. So, for example, you can do demand forecasting on as little as a few weeks of historical data.
Now, to get back to your question. How can one capitalize on this knowledge? I'd say, first of all, definitely invest your time to educate yourselves, whether by taking classes, talking to experts, or reading related materials. Don’t wait for the future to come upon you and take you by surprise. Keep educating yourself to stay ahead of the curve. If you want to stay relevant in the next 10 years, you need to learn about AI.
That’s fair, but one could argue that it’s easier said than done. Not everyone in the business world has a statistic background or a math degree or is even math-savvy, for that matter. Even basic statistics principles can be challenging to grasp, not to mention those more advanced algorithms. The latter not only requires a good comprehension of the science behind it but also requires programming skills, such as Python or R, to be used for problem-solving.
Excellent point! As I said earlier, these algorithms used to be a prerogative of scientists just a few years ago. But things change fast, and the tendency for skills and knowledge democratization touched the field of AI as well. So it’s important to distinguish between being able to build a tool versus being able to use a tool.
A common example of knowledge democratization is, of course, Google. How often do you use it to find the correct spelling, self-diagnose based on your symptoms, or book a hotel room? Traditionally, those functions were exclusively limited to trained individuals. And now, these are accessible to anyone with a network connection and the simplest smartphone.
Consider the example of a car. In order to drive it, you don’t need to know the mechanics of it or understand all the forces that power its movement, the principles of the internal combustion engine, etc. You just need to learn to drive—a skill that kids legally qualify for earlier than they graduate from high school.
Similarly, you don’t need to code to use AI nowadays, nor do you need a data science degree to solve complex computational problems. Just think about this: A hundred years ago, only a few elite mathematicians could perform logarithmic functions, then calculators made this available to anybody. The next iteration of this came with a computer—"a bicycle for the mind”—and its spreadsheets that enabled data analysts to form more advanced computations, such as DCF analysis or linear regressions.
And now, we have entered the next phase of upgrading our computational skills—AI-powered analytics. These no-code AI tools enable anybody to be a so-called citizen data scientist. This means you can use sophisticated algorithms to solve business problems that otherwise cannot be tackled. For example, try comparing 10^19 scenarios to find the most optimal one. It would take you a lifetime to solve this with spreadsheets or hundreds of lifetimes to solve it with a pen and paper (this is what I call the multiverse of scenario analysis madness). With AI, you can solve it in minutes.
Why do we need to get so deep into data analytics? It sounds like we’ve been doing just fine with simple spreadsheet-based analysis. Some say it can wait until next year – or later.
There were stores that, 20 years ago, claimed they did just fine with only physical locations and did not need e-commerce. I think we all know what happened to them.
But if we were to look at some tangible and short-term impacts, it already makes sense to act upon this opportunity fast. Consider this: If you make $50M in sales and you under-forecast by 1%, $0.5M is missed revenue.
And if we were to get back to your original question about future-proofing one’s career, consider this: As the adoption of spreadsheets is transitioning from a mature to a declining phase, these skills are getting more and more commoditized. So, to differentiate yourself from your peers in the short term and to prepare for the next bump in computational complexity in the long term, you need to embrace AI. The easiest way to do it is through the no-code tools out there. If you’re not at least experimenting with AI-driven analytics, you’re already behind.
Not sure how to get started with no-code AI? We got you!