AI and Your Work: What Can You Do So You Don’t Stay Stuck in Place?

Do you expect a day to come when AI takes your place at work?

Picture this: part of your workday, the reports, the messages or the designs, has started to involve AI.

The question here isn’t: “Will these parts change?” They already are.

It’s: “How will you change, and adapt to them?”

When the internet first appeared, many people feared it would take their place. What actually happened is that those who didn’t adapt fell behind, and those who learned it changed the way they worked and became stronger in the market.

Today AI is bringing the same idea back, only faster and more complex. Some jobs will be affected, some tasks will disappear, and new roles will be created.

So when someone says, “AI will replace humans,” the more accurate answer is: it will change the way we work, yes, but it will favor people who know how to work with it, not against it.

A Job Isn’t One Thing

When we ask, “Will AI take my job?” we picture the job as a single block: it either stays or it disappears.

But economists see it differently: a job is a bundle of tasks.

In a well-known 2015 article titled “Why Are There Still So Many Jobs?”, MIT economist David Autor explains that automation does replace people in routine tasks that can be written down as rules, but at the same time it complements their work in other tasks and raises output, which increases the demand for human labor.

He argues that commentators tend to overstate the replacement and overlook this complementarity.

The recent picture says something similar. In an updated global index released by the International Labour Organization in May 2025, about one in four workers is in an occupation with some degree of exposure to generative AI, while only 3.3% of jobs fall into the highest exposure category, most of them clerical work.

The organization’s conclusion: because most occupations include tasks that need a human, transformation is the more likely effect, not full replacement.

What Does the Research Say So Far?

The research is still in its early days, but three studies give a balanced picture:

• It raises performance, especially for beginners: In a study of 5,179 customer service agents, an AI assistant increased the number of issues resolved per hour by about 14% on average, and by about 34% for new and less-skilled agents, while its effect on experienced agents was minimal.
• Excellent inside its limits, misleading outside them: In an experiment with 758 employees at a global consulting firm, those who used it completed 12.2% more tasks, 25.1% faster, and at higher quality. But on a task deliberately chosen to fall outside its capabilities, they were less likely to reach the right answer than those who worked without it. The researchers call this the “jagged frontier”: you don’t always know where the tool’s ability ends.
• The first door into work is changing: A Stanford University study from August 2025, based on U.S. payroll data, found that employment of young people aged 22 to 25 in the occupations most exposed to AI fell by about 13% compared with others, while employment of experienced workers in the same occupations held steady or grew.

Each of these studies has its limits: one market, one sector, a short period. But together they say the effect is real, and uneven.

Where Does This Show Up in Everyday Life?

Today there are tools that help with programming, writing, design and marketing. Not to cancel out the human, but to free people to focus on what they’re best at: thinking, deciding and innovating.

An accountant at a company in Baghdad uses it to summarize the monthly expense sheets in minutes. But she checks every number before the report reaches her manager, because she’s the one who’ll be asked about it.

A physics teacher uses it to prepare practice questions at different levels, then picks the ones that suit his students and edits the rest. The tool wrote them, but he’s the one who knows his class.

And a young man who works in marketing tried it once with a single line, got back a generic, lifeless text, and said: “It’s useless.”

The problem usually isn’t the tool. It’s the request. The company behind Claude advises in its guide to treat it like a very smart new employee who knows nothing about how you work: the more precisely you explain what you want, the better the result.

The Common Mistake: “Learn the Tool and You’ll Be Safe”

Here we need to be honest.

Learning AI is not a guarantee of complete safety.

• Tools change fast: What you learn today may look different within a year. What lasts is your understanding of your field, and your ability to judge the result.
• The impact doesn’t hit everyone equally: The latest study suggests the first people affected may be those at the start of their careers, not only those who ignore the tool.
• Blind trust is a weakness too: The “jagged frontier” experiment says that someone who hands their mind over to the tool may make more mistakes than someone who doesn’t use it at all.

I’m not saying everyone who doesn’t learn AI will be replaced. But whoever ignores it will become less competitive over time.

Because AI amplifies the abilities of those who understand it, and exposes the weakness of those who ignore it… and it also exposes the weakness of those who trust it without checking.

How Do You Start, Calmly?

You don’t need a long course to begin. Try these steps over the next two weeks:

1. Map your tasks: Write down the tasks of a typical week, and next to each one note whether it’s repetitive, needs judgment, or needs a relationship with people. Example: “Entering invoice data” is repetitive, while “Persuading an angry customer” is about relationships.
2. Start with one repetitive task: Try a tool that serves your field on it for two weeks. Example: Summarizing meeting minutes instead of writing them from scratch.
3. Write your request as if you’re talking to a new colleague: The role, the context, the request and the format. Example: “You’re an editor. This is this month’s sales report. Summarize it for the manager in five points, in simple language.”
4. Check everything: Numbers, names and sources. Example: If the answer mentions a study, look it up yourself before you pass it on.
5. Invest in what can’t easily be copied: Your deep understanding of your field, your relationships, and your ability to make decisions. Example: The tool writes the proposal, but you’re the one who understands what’s worrying the customer.
   

A Simpler Mental Model

• A job is a bundle of tasks, and AI changes some of them, not all.
• It boosts the performance of those who work with it, and hurts those who trust it without checking.
• Learning it isn’t a guarantee, but ignoring it is a real risk.

Start calmly: learn the basics, try the tools that serve your field, and see how you can make them raise your value in the market, not replace you.

Think of it this way:

The question isn’t “Will AI take my place?”

It’s: What will I do so I don’t stay stuck in place?

Today, pick one repetitive task from your work, try doing it with AI, then review the result with a critical eye.

Further Reading

• Why Are There Still So Many Jobs? The History and Future of Workplace Automation — David H. Autor, Journal of Economic Perspectives, 2015: Why automation hasn’t ended most jobs throughout history, and how it replaces some tasks while complementing others.
• Generative AI and Jobs: A Refined Global Index of Occupational Exposure — International Labour Organization, 2025: A global index showing which occupations are most exposed to generative AI, and why transformation is more likely than replacement.
• Generative AI at Work — Brynjolfsson, Li & Raymond, NBER, 2023: A field study of customer service agents showing who benefited most from an AI assistant.
• Navigating the Jagged Technological Frontier — Dell'Acqua et al., Harvard Business School, 2023: An experiment showing when AI improves work and when it leads to mistakes.
• Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence — Brynjolfsson, Chandar & Chen, Stanford, 2025: What U.S. payroll data says about AI’s effect on hiring young people at the start of their careers.
• Prompting best practices — Anthropic, Claude Docs: A practical guide to writing clearer requests, including the advice to treat the model like a smart new employee.