Is AI Really Going to Take Your Job?
For technology leaders, the key issue is not whether AI removes jobs, but which tasks, decisions and accountabilities move up or down the stack. That shifts the management question from headcount reduction to role redesign. CIOs and programme leaders should treat AI as a workflow and capability change: what gets automated, what gets reviewed by humans, and where new oversight, quality assurance and exception handling are required.
That has direct implications for operating model design. If AI makes routine drafting, coding or analysis cheaper, demand can expand faster than expected, but only if teams can absorb the extra output. The trade-off is clear: higher throughput and faster response times on one side; more coordination, review burden and risk of inconsistent use on the other. Leaders need explicit decision rights for where AI is allowed to act, where it must recommend, and where humans must approve.
The next management question is capability, not just adoption. Upskilling cannot be limited to prompt writing. It should include model oversight, validation, incident escalation, and judgment in ambiguous cases. This is especially important for portfolio leaders deciding whether AI should be funded as a productivity layer across functions or as a set of narrowly governed use cases. The second route is safer; the first is more scalable.
Practical questions for executives include: Which roles will be augmented first? What measures will prove value beyond anecdotal productivity? How will managers detect over-reliance, quality drift or hidden rework? Organisations that answer these questions early are more likely to capture AI-enabled growth without turning efficiency gains into control problems.
No, but your role is probably going to look a little different.
Loading the Elevenlabs Text to Speech AudioNative Player…
Whether itโs fielding basic customer complaints or generating sophisticated code, it seems like todayโs AI tools can just about do it all. But as the technology keeps growing at a rapid clip, so have employeesโ fears over its ability toย make them redundant.
These concerns arenโt unfounded. Companies including Microsoft, Amazon and Salesforce have cited AI adoption as the reason behindย over 50,000 job cuts in the United States in 2025. More recently,ย The Wall Street Journal reported that AI companies are asking subject-matter experts toย train their models in everything from astronomy and psychology to video editing and financial investing. OpenAI has also apparently been asking contractors toย upload projects from other jobs to train and prepare its AI agents for office work.
Will humans be left in the dust as AI takes over? Iโd argue that history and simple economic analysis suggest otherwise, and that conventional wisdom about the technology annihilating jobs is almost certainly wrong.
A look at past technological advances
Letโs take it back to the 1980s, when CT scans became appreciably cheaper thanks to advances in medical imaging technology. The worry was that everyone employed in radiology departments would lose their jobs. But per the US Bureau of Labor Statistics, employment in this area didnโt fall โ it actuallyย grew, leading to aย Original Postii/S1546144004002285" shape="rect">shortage of radiology technicians.
What happened was that more people could afford getting CT scans. More doctors recommended them, and insurance companies were more willing to pay for them. The resulting efficiency led to an uptick in CT scans, which created jobs, not killed them. It increased the need for human oversight, unearthed more complex cases (that subsequently needed treatment) and led to the emergence of new specialties like interventional radiology.ย
Another example is transportation. Ride-sharing apps like Uber commodified mobility when they first burst onto the scene in the late 2000s, dramatically dropping the cost of a ride. Youโd expect that many drivers, especially taxi drivers, would lose their jobs. Instead, lower costs exploded demand. People who typically walked or took buses started hailing rides for short trips, while new sectors like food delivery emerged. Globally, the number of ride-hailing drivers ballooned from thousands to millions. Sure, some taxi drivers had to become Uber drivers, butย overall employment in the ride-sharing and taxi industry grew substantially between 2015 and 2023.
We can go back in history to the Industrial Revolution to see the same pattern. Until the 1870s, prior to the invention of something called theย Bessemer process, buildings were limited to materials like wood, brick and iron, which couldn’t support extreme heights or withstand environmental stressors without costs becoming prohibitive.ย
The Bessemer process of producing steel cost-effectively changed that, allowing us to build much taller buildings. But it also led to fears over job displacement in traditional ironworking trades, which manifested in strikes, union resistance and shutdowns. Although there were some shifts in labour types โ and steel production reduced the need for highly skilled ironworkers โ new jobs sprouted and demand for overall construction labour increased.
Jevons paradox: efficiency in action
The idea behind this, known as Jevons paradox, was first articulated in 1865 by economist William Stanley Jevons. He noticed something about the use of coal in England: New steam engines made coal use more efficient, but instead of reducing the amount of coal required, it made coal cheaper and more versatile. Industries ramped up their consumption, deploying it for everything from powering factories and ships to heating homes. Demand for coal skyrocketed, and so did jobs.
Letโs apply this logic to AI and knowledge work. AI agents make non-routine tasks like coding prototypes, drafting reports and brainstorming campaigns cheaper and more accessible. This means that a small start-up can use AI to avoid shelling out significant resources on bespoke software, allowing it to tackle more projects, build more apps and input more features. Demand for coders doesn’t drop. It rises because customised software is being used by more companies. Beyond writing code, coders will now do different things, such as testing code generated by AI, as well as providing oversight, integration and advanced tweaks.
You can see a similar trend in the graphic design industry. Prior to the invention of AI, graphic design was time- and cost-intensive. Now, companies can create ads quickly and easily. Graphic designers have to use AI to generate more ads within a shorter timeframe, but the need for human collaboration, good taste and a strong connection to the brand and audience will remain โ things that AI has yet to fully replicate.
Pre-AI, you often had to shell out a substantial fee for basic legal documents. With AI making legal work cheaper, the demand for it will increase, and lawyers will have to assume a more advisory role rather than just executing legal documents. The same goes for mergers and acquisitions, where AI has brought down the cost of deal-making. Bankers will then act as advisers โ not so much as writers of contracts and covenants. There will be more deals happening, and more work for bankers, but in a different capacity.
How to give yourself an edge
So, take heart. All is not lost, not by a long shot.ย
Thatโs not to say that the transition to an AI-infused world will be easy, or that there wonโt be pain in the form of layoffs and reskilling. Some may gripe that a few companies and individuals seem to beย reaping the spoilsย of AI, while the rest of the population is left having to adjust to a new technology that, frankly, many of us didnโt ask for.
But what’s clear is that AI will likely affect every job in one way or another, and employees must adapt as companies increasingly incorporate the technology into how they do business. McKinsey, for instance, is piloting anย AI skills testย as part of its recruitment process by asking applicants to use the firmโs AI tool, Lilli, to complete certain tasks. Candidates are evaluated on how they prompt the chatbot and what they do with the information it generates.
How can you ensure that you gain from AI? Here are some easy-to-implement tips:
- Embrace it as a multiplier: Use it for drafts and ideation, then add your human edge. By using AI to improve your performance and get work done better and faster, you can free up time for higher-level tasks.ย
- Upskill in oversight: Learn how to manage AI agents โ which can be extremely useful if you use them well โ such as by taking courses on prompt engineering specific to your job or industry. You can even ask AI agents to summarise the latest knowledge about how to make the best use of AI for your context.
- Spot opportunities: Ask yourself, โWhat in my field will be unlocked if this task gets 10 times cheaper?โ Capture these opportunities, either in your existing job or by becoming an entrepreneur. Indeed, itโs not difficult to create custom AI agents that specialise in a particular job function.
- Lead on AI adoption: If youโre running a team, normalise AI in workflows. Use it to share knowledge more quickly and solicit ideas from your employees on how to better use it.
Edited by:
Rachel Eva LimAbout the author(s)
Nadav Klein
is an Associate Professor of Organisational Behaviour at INSEAD. His research focuses on the basic processes of judgment that affect how people make decisions, process information, and evaluate others and themselves.
Enjoyed this article? Sign up for our newsletter to receive regular insights and stay connected.