The work does not disappear. It moves. Data professionals still need to determine whether the problem is meaningful, the data is appropriate, the output can be trusted and an AI-generated prototype can move safely into a production environment.

Data professionals increasingly need to ask:

• Is this the right problem to solve?
• Is the underlying data suitable for the task?
• Can the output be trusted?
• What happens if the model is wrong?

AI can accelerate the first draft. Human expertise determines whether that draft deserves to become a decision, a product or a production system.

Are data analyst and entry-level roles at risk?

Some routine tasks are more exposed to automation than others.

AI can already help with basic reporting, starter code and simple dashboards, which may put more pressure on roles built primarily around production.

That does not mean data analysts or entry-level professionals are becoming obsolete. These job titles have always covered a wide range of work, from preparing recurring reports to helping leaders interpret evidence and make decisions.

Data quality is another reason replacement is not so simple. Many organizations have powerful AI tools but still lack the clean, reliable data needed to use them well. They need professionals who understand where the data came from, whether it can be trusted and how it should be applied.

AI may change the first steps of a data career, but it does not remove the need for human judgment. The opportunity is for professionals to use AI to move beyond routine production sooner and contribute more directly to analysis and decision-making.

Technical foundations matter more than ever

The rise of AI coding assistants might suggest that technical knowledge is becoming less important. In practice, the opposite may be true.

The career trajectory of Mohamed A. Mneimneh, an adjunct lecturer for Brown’s online data science master's program, shows how widely machine learning can be applied. His graduate research ranged from predicting failures in Navy motors and generators to developing models for earlier cardiac disease detection, a focus shaped in part by his father’s experience with heart disease.

“When AI writes the code, it does not mean the human no longer needs to understand it,” Mneimneh said. “Someone must still examine the process, audit the logic and determine whether the code reflects the intended problem.”

AI-generated code is a starting point, not a finished product. The first draft may arrive quickly, but testing, debugging and integration still require technical fluency.

Several foundational capabilities are becoming especially important:

• Mathematics, statistics and probability
• Programming logic and code auditing
• Systems design and data engineering
• Debugging, deployment and production readiness

The goal is not for data scientists to compete with AI on speed. It is to know enough to direct, question and improve what it produces.

Governance is becoming part of the technical job

As organizations move from experimenting with AI to using it in real workflows, governance is becoming part of the technical decision-making process.

Data professionals need to consider where bias may exist, whether privacy and compliance requirements are being met and whether the data, access and outputs are appropriate for real-world use.

Cost is part of that judgment, too. Teams need to understand what data is being shared, what is actually necessary and whether the value of using an AI tool justifies the cost and risk.

Responsible AI is not about using a tool simply because it is available. It is about knowing when to use it, how to use it and when a different approach may be better for the organization.

Asking the right question is becoming a defining skill

AI can answer a question without understanding whether it is the right question. 

A model may have access to a dataset without understanding how the information was collected, why certain fields are missing or which constraints matter to the organization using it. It can produce a technically plausible answer that has little relevance to the decision at hand.

That is where professional judgment becomes especially important. Professionals who understand an organization’s goals, data, customers and constraints can frame better questions before an analysis begins.

As Mneimneh puts it: 

“A technically strong solution still falls short if it is not tied to the right problem.”

Strong problem framing means asking:

• What decision are we trying to support?
• Which assumptions are built into the data?
• Whose perspective may be missing?
• What could happen if the system is wrong?

This kind of judgment is difficult to automate — and increasingly important as AI makes technical output easier to generate.

Why storytelling and other soft skills matter in data science 

Data science does not end when an analysis is complete. Professionals also need to explain what they found and why it matters.

A technically strong result has limited value if stakeholders cannot understand what it means or how to act on it. Data scientists need to communicate complex findings into a clear story that connects the work to a real decision.

Firrincieli describes this as an essential part of turning technical work into impact:

“You can produce strong analysis and meaningful results, but the real question is whether you can turn them into an understandable, compelling story that business stakeholders can act on. In many work settings, being able to communicate your findings clearly is just as important as the technical work itself.” — David Firrincieli

Not everyone contributes to data science in the same way. 

One person may focus on machine learning while another brings deeper expertise in visualization, communication, policy or a particular industry. Those skills play a real role in the technical process. They help determine whether the results are understood, trusted and used.

What AI skills are employers looking for?

Employers are increasingly looking for evidence that candidates can use AI thoughtfully in real work, not just name the latest tools.

Strong candidates can show where AI added value, how they evaluated the output and when they chose not to rely on it. The focus is less on communication itself and more on judgment, application and results.

An example should show:

• The problem AI was used to solve
• Why AI was the right approach
• How the output was checked
• What changed or improved
• Where human judgment was still needed
• What risks, limits or tradeoffs had to be considered

While platforms such as ChatGPT, Claude, Gemini, Codex and Claude Code will continue to change, the more durable skill is knowing how to use AI effectively without handing over professional judgment to the tool.

AI is creating new career paths while changing the shape of existing ones 

Data science is drawing on a wider range of expertise as AI changes how the work gets done. Technical skills still matter, but industry knowledge, policy experience, communication and an understanding of how technology affects people and organizations are becoming increasingly valuable. 

Linda Clark, associate teaching professor of data science and academic director of Brown’s online master’s program, Data Science: Policy, Governance and Society (DSIO), describes this broader perspective as transdisciplinary.

For professionals, that means moving into data science does not always require leaving previous experience behind. Expertise in health care, finance, retail, government or another field can become an advantage when paired with stronger technical capabilities.

What new jobs are emerging in data science because of AI?

AI-related hiring is already showing significant momentum. LinkedIn reports that U.S. AI job postings have roughly doubled since 2023, with AI engineer ranked as the fastest-growing job in 2025. 

At the same time, organizations are building new leadership, governance and implementation functions around AI, creating opportunities well beyond traditional data science roles. 

• AI engineer

A strong fit for software, machine learning or engineering backgrounds. AI engineers build and test applications using machine learning models and large language models.

• AI solutions architect

Often suited to professionals in systems, cloud, data engineering or enterprise technology. They design the infrastructure and systems that support AI applications.

• AI governance and risk lead

A natural fit for risk, compliance, privacy or governance backgrounds. These professionals help organizations manage AI risk and responsible use.

• AI product manager

Product, business and operations experience can translate well here. AI product managers connect user needs, business goals and technical capabilities.

• AI policy and legal professional

Legal and policy backgrounds are increasingly valuable as organizations navigate AI regulation, accountability and governance.

• Technical leader and AI strategist

Experienced technical or business leaders may focus on evaluating AI opportunities, setting direction and deciding where AI can create value.

These paths show that professionals do not necessarily need to start over. In many cases, the opportunity is to combine existing expertise with stronger data and AI skills.

What should you look for in a master’s in data science program?

Data science programs should not be defined by whether they teach the newest AI platform. Tools change too quickly for that to be a meaningful measure on its own.

A better question is whether a program is designed to help students adapt as the field changes.

That includes:

• A focus on lasting skills, not just today’s tools
• Real projects with messy, real-world data
• Practice knowing when AI helps and when it doesn’t
• Exposure to business, ethics and policy considerations
• Opportunities to explain and apply technical work
• Perspectives from different industries and disciplines
• A curriculum that keeps pace as the field changes

Earn your master’s in data science at Brown

Brown University’s online master of science in data science: policy, governance and society, offered through Brown University’s Data Science Institute, reflects the broader direction of the field.

The program combines technical training in areas such as machine learning and data engineering with policy, governance and social impact. Through live sessions, collaborative learning and applied projects, students work with real-world data while developing the judgment, communication and decision-making skills needed to move from analysis to action.

As AI continues to reshape data science, the program helps professionals strengthen their technical capabilities while building the transdisciplinary perspective needed to evaluate problems, guide responsible use and lead in an AI-driven environment.

Take the next step in shaping the future of responsible AI.