Ask an MIT Professor: Why Data Science Skills Matter in the Age of AI

What value do foundational skills in data science and statistics hold in a world dominated by AI and automation?

The answer might surprise you. 

As advanced technologies become more integrated into our lives and work, experts like MIT Professor Devavrat Shah posit that comprehending the mathematical principles behind them is still critically important. 

Professor Shah is Faculty Director of the MIT MicroMasters Program in Statistics and Data Science, a multi-track program designed to provide this foundational knowledge. 

With advanced tools available, why bother with foundational data science skills?

Professor Shah takes a broad view when considering the value of foundational data science skills in the age of advanced technology. He likens the question of whether we still need to learn basic skills to asking, "If we've already built all the bridges, why would anyone need to study civil engineering? If we've mastered electricity transmission, why do we still train power engineers?"

He draws a parallel with calculators, questioning, “Why teach basic math when we have tools that can do it for us?” The answer, he explains, is clear: foundational skills remain essential. 

We still need civil engineers to design and build new bridges and power engineers to innovate and maintain our electrical grids. Children around the world continue to learn math in school because these fundamental concepts are vital, even in a world dominated by advanced tools and technologies.

In the same way, Professor Shah acknowledges that generative AI automates specific tasks, but effectively using this powerful tool requires a deep understanding of where and how to apply it.

“This is why foundational knowledge, such as what is taught at institutions like MIT, remains crucial,” he says. “Just as we need engineers who understand the principles behind bridges and power grids, we need individuals who can grasp the underlying concepts behind AI to truly leverage its potential.”

He adds that a key element of learning is the ability to “tinker.” While automated tools can be efficient in a work environment, there is little room for experimentation. Focusing on foundational knowledge gives learners a “sandbox” where they can tinker and experiment to deepen their understanding. 

How does the evolution of technology influence the focus of academic content? 

Professor Shah observes that academic content naturally evolves as technology advances. 

He uses circuit design as an example. There was a time when electrical engineers and computer scientists spent much of their work learning the mechanics of circuit design. As circuit design became increasingly automated, there was less need for human intervention. However, technology companies need skilled circuit designers to create specialized circuits for modern devices like smartphones. The demand for expertise in specific, high-impact applications is high. 

“In that sense, we’ve moved away from teaching some of the nitty-gritty details of circuit design, but we still cover certain aspects,” says Professor Shah. “What we used to teach has been replaced by new content. Over the last 10 to 20 years, much of the teaching in computer science departments has focused on helping people think through how to build AI infrastructure.” 

He continues, “As AI infrastructure becomes more consolidated and industry standards emerge, perhaps that’s not what we may end up teaching in the years to come. Maybe we should focus on something else. There will be an evolution, but at the same time, some core elements will remain. For example, we still teach arithmetic, which has endured for thousands of years. Some things are timeless.”

Why is it important to learn foundational concepts in probability and statistics in data science, even with AI tools available?

Professor Shah argues that despite AI's advancements, it has inherent limitations that can lead to significant oversights if users don't grasp the fundamental concepts behind it. He illustrates this idea by highlighting the probabilistic nature of AI and how it contrasts with the deterministic models many of us are accustomed to. 

When AI models predict outcomes, they do so based on probabilities rather than certainties. Understanding how to interpret these probabilities is crucial to making informed decisions. Without this nuanced understanding, one might misinterpret AI outputs, leading to flawed decisions or misplaced confidence in the results.

Take weather forecasts as an analogy. When you hear that there’s a 60% chance of rain, it does not simply mean a 60% probability, like flipping a coin. This percentage reflects a complex model that predicts rain in 60% of a given area. 

However, the real challenge lies in understanding what probability means in practical terms. It’s not just about the likelihood of rain; it’s also about your confidence in that prediction. If we don't understand these probabilistic concepts, we risk making poor decisions based on AI's predictions and treating them as deterministic when they are not.

The same principles apply to AI in data science. An AI tool might give you a prediction, but without a fundamental understanding of probability and statistics, you might not fully comprehend the confidence level or the potential variance in those predictions. This misstep could lead to significant errors in judgment, especially in fields where precise decision-making is critical. 

Therefore, learning foundational data science skills is essential—not only to understand AI's outputs but also to use those outputs effectively and responsibly in real-world applications.

How can the MicroMasters Program in Statistics and Data Science help learners in their professional lives?

Professor Shah highlights that many of us now work in large enterprises that rely on complex departmental and interdepartmental workflows. Decades ago, these workflows were paper-based, with information manually recorded and shared. Over time, digitization transformed these processes, followed by a shift to cloud-based systems. 

Now, we're entering a new phase—what he calls “AI infusion” or “AI migration”—where AI is becoming deeply integrated into the processes. The MIT MicroMasters Program in Statistics and Data Science will equip learners with the skills needed to navigate this landscape. 

Whether you're a leader driving AI adoption, a manager overseeing its implementation, or a contributor directly involved in using these tools, this program provides the foundational knowledge you need to deliver value.

Prepare Yourself to Lead in an AI-Driven Future

As AI continues to reshape industries, the demand for professionals who can navigate and harness these technologies is only growing. However, understanding the fundamental concepts behind AI tools is just as crucial as knowing how to use them. This knowledge is what enables data scientists to make informed, responsible decisions that drive real impact.

The MIT MicroMasters Program in Statistics and Data Science is designed to prepare you for this challenge. By building a strong foundation in probability, statistics, and data science, you’ll be equipped to lead in a world where technology and data are central to success. 

Enroll in the program today and position yourself at the forefront of this evolving landscape.