Historically, technology has always shaped—and disrupted—labor markets. Steam engines and computers pushed entire professions into extinction, from telegraph operators to elevator attendants. Now, generative AI is stepping up as the next “general-purpose technology,” with nearly 40% of Americans already using AI tools at home or at work. That adoption rate outpaces both the PC and the internet in their early days.

Instead of speculation, Microsoft’s research focuses on what AI is currently doing. The data collected over nine months in 2024 from users in the U.S. highlights how AI is already performing tasks across a wide range of professions.

Among the most common tasks users sought help with were information gathering and writing. Think office assistant, but digital and highly scalable. But there’s more to it: the study emphasizes the distinction between the user’s goal and the AI’s action. For instance, a user trying to print a document has the goal of “operating office equipment,” while the AI’s action is to “instruct use of equipment.” This subtle difference helps identify whether AI is completing the task or just guiding the user.

To quantify AI’s impact, Microsoft developed an “AI applicability score” for each profession, combining usage frequency, task success and potential influence. It used the O*NET database, which breaks jobs into hierarchical work activities.

So who’s most at risk?
Jobs involving knowledge work—such as IT, mathematics and office support—show the highest AI applicability. Surprisingly, sales roles also rank high due to the focus on communication and information delivery. But topping the list, by a wide margin, was translation: 98% of a translator’s tasks overlap with common Copilot queries. Other vulnerable professions include historians, writers, CNC programmers, brokerage clerks, political scientists, journalists, mathematicians, proofreaders, and editors.

Conversely, roles requiring physical human interaction, such as nursing assistants and massage therapists, or manual labor (like dishwashers, construction workers and truck drivers) are far less affected, at least for now.

Will AI replace workers—or empower them?
The age-old question remains: will AI eliminate jobs or enhance them? Microsoft’s study doesn’t take a definitive stance. “The data focuses on usage, not outcomes,” researchers caution.

Looking back at ATMs, for example, they replaced one core bank teller task but led to a rise in teller jobs overall by freeing up staff to do higher-value work. Likewise, the study suggests AI is acting more as a service—a tool for training, consulting, or simplifying information—rather than outright replacing humans.

Information gathering and writing not only rank as the most common AI uses, but also receive the most positive user feedback. Yet despite AI’s effectiveness, the report notes limitations: an AI-written novel might struggle to emotionally engage readers, lacking lived human experience.

Interestingly, the study found a weak correlation between AI applicability and salary—meaning high-paying jobs aren’t necessarily more exposed. However, jobs that require a bachelor’s degree tend to score higher on the AI applicability scale than those requiring less education.

Could AI usher in the proletariat revolution Karl Marx imagined? It’s too early to say, but Microsoft’s data does hint at major workforce changes.

A rare glimpse into a historic shift

While the research is limited to one AI platform and a specific dataset, it offers a rare real-time snapshot of labor market transformation. What was once science fiction is now laid out in charts and graphs.

Although the data focuses on the U.S., the trends are expected to affect most developed and developing countries—especially in the absence of global AI regulation. Even Microsoft has laid off thousands this year, citing AI as a reason for reshaping internal workflows.

At the safer end of the spectrum, the study lists 40 professions less likely to be affected—mostly those involving physical tasks, fine motor skills, or complex human judgment. Think physical therapists, construction workers, farmers, engineers and surgeons. However, as robotics continues to evolve, even these roles may face disruption down the line.

In the end, the biggest concern is that AI could deepen inequality and fuel political instability, prompting populist shifts. On the flip side, the technology also holds tremendous promise—such as speeding up cures for diseases like cancer by analyzing data far faster than human researchers.

History shows that every industrial revolution has caused upheaval, but also expanded opportunities, industries and life expectancy. The big question remains: will AI create enough new jobs to replace the ones it’s displacing—or are we heading toward a world where work itself becomes obsolete?