Teleporting to the Mountaintop: The Emotional Cost of AI at Work

Alex did not expect his work life to turn into a series of teleports.

For most of his career, big projects felt like long hikes. You started at the trailhead with a vague map and a list of deliverables. You gathered requirements, walked through meetings, stepped over bottlenecks, stopped to revise, and eventually, if the weather held, you reached a summit that looked like a finished report or a working prototype. It took weeks. There was time to second-guess a decision, to sleep on a tricky paragraph, to notice if something felt wrong.

Then his company switched on its new AI tools.

The first time Alex tried them on a real piece of work, the ground moved. Tasks that had lived quietly in the “several days” category collapsed into hours. A messy document became a clean summary in minutes. A first draft of a client memo appeared in the correct tone while he was still sipping coffee. A project plan that once demanded a full afternoon surfaced in a single, slightly uncanny outline.

By Tuesday, he was standing on a Friday summit.

Across town, Rosa was still on the trail.

She worked at a regional insurer, processing claims and answering calls. Her days were structured by queues and scripts: a claim to review, a customer to calm, a supervisor to consult. When the company announced a “virtual assistant” to handle simple inquiries, she read the memo, clicked through the training, and went back to her headset. Software came and went. The work stayed the same.

Over the next few years, Alex and Rosa walked into the same technological future from opposite sides. For Alex, AI became a kind of exoskeleton at work. It let him lift ten projects instead of two. For Rosa, AI arrived mainly as something that routed her calls, scored her performance, and determined her schedule. The tasks changed; the power did not.

On paper, they were employees of companies “embracing AI.” In reality, they were inhabiting two different stories about what AI does to a life.

This article follows those stories and the economy they point to: how AI is changing the shape of work, what happens to a consumer society when workers like Alex and Rosa are pulled apart, and what it feels like to live with a system that can teleport you to the mountaintop before you’ve had time to catch your breath.

What AI Is Actually Doing to Work

Beneath the marketing language, three simple changes explain much of what Alex and Rosa are experiencing.

1. Jobs are bundles of tasks, and AI works at the seams

A job title sounds singular: nurse, claims processor, product manager. In practice, each job is a bundle of tasks.

A nurse charts notes, adjusts medications, comforts families, coordinates with doctors, and handles the quiet logistics of a ward. An accountant tracks receipts, reconciles numbers, writes emails, meets clients, and interprets rules. Rosa, in her call center, answers questions, updates systems, de-escalates angry customers, and knows which supervisor is best for a particular mess.

AI rarely replaces an entire bundle in one bite. It goes after certain kinds of tasks: the ones that are repetitive, pattern-based, heavily textual or numerical. Summarizing. Drafting. Sorting. Answering standard questions.

For Alex, this is immediately visible. The glue work that used to soak up his day collapses. Status reports, first drafts, data cleaning, slide skeletons: they can be handed off to a system that works at machine speed. His job doesn’t disappear; its composition changes. He is pulled closer to the edge of his role where judgment and negotiation live.

For Rosa, the same slicing shows up differently. The FAQ-style questions peel away to the chatbot. What remains are the difficult calls: people who are already frustrated, whose problem didn’t fit the template, who are asking for judgment, exceptions, or simply someone to listen. The soothing, predictable rhythm of simple calls thins out. The emotional load thickens.

Nothing in the technology guarantees one experience or the other. But the seams of jobs are exactly where AI starts to press.

2. One person can move a lot more weight

The second shift is easier to feel than to see on a spreadsheet. AI acts like a workplace exoskeleton.

Put it on, and your output changes. A single doctor with a well-designed digital system can see more patients safely. A single engineer with powerful tooling can ship more code. A single coordinator with a capable assistant, even a synthetic one, can manage more projects.

For Alex, the exoskeleton is exhilarating. When his company rolls out a large language model internally, he starts experimenting. He asks it to extract action items from meeting notes. That works. He asks it to suggest risk scenarios. Those are usable. He feeds it fragments of old proposals and asks for a draft in house style. It’s not perfect, but it’s far enough along that he can fix it instead of starting from zero.

Within months, the number of things he can “handle” doubles. He becomes the person who never seems to miss a deadline. Managers notice. When a new initiative appears and someone asks, “Who can take this on?” his name is on the short list.

There is a visible story here: a smart, curious worker using a new tool well. There is also an invisible one: each success becomes an expectation. If you can lift ten boxes, the room quietly reorganizes around that fact. Soon, the question is not “Isn’t it amazing that you can do this?” but “Why aren’t all ten slots on your calendar filled?”

Rosa’s exoskeleton is attached somewhere else. The AI that first “helps” by answering simple customer queries soon sits in the routing engine and the quality-assurance dashboard. It determines which calls she receives, when they arrive, how her tone is scored, whether she stayed within the scripted boundaries. It is very good at counting.

Her effort is measured more precisely. Her breaks become more visible. The system can recommend “coaching modules” based on the shape of her voice in the last hour. To the company, this is efficiency and consistency. To Rosa, it feels like working for a spreadsheet that watches back.

For both workers, AI increases the amount of weight that can be moved in a day. The difference is who is allowed to steer and who is simply braced under the load.

3. The gains do not distribute themselves

When productivity rises, someone benefits. That is almost the only guarantee.

The extra value created by AI can go into higher wages, lower prices, greater profits, more hiring, shorter workweeks, better services, or some mix of these. The technology does not choose. Institutions do.

If Alex’s company sees his new capacity as a sign of leadership potential, and if it has norms that tie performance to pay and promotions, he may feel some of the gains in his own life. If the culture treats increased capacity mostly as a way to “do more with less,” he may simply find himself doing the work of multiple people for a slightly larger salary and a more nervous smile.

If Rosa’s company decides to use AI-generated insights to lighten the heaviest parts of the job, adjust staffing, and ensure more sustainable workloads, she might experience the tools as support. If the company’s main aim is to contain costs and increase call volume per agent, the same tools tilt toward pressure.

This is where the individual stories touch the larger economy. A million small decisions about where AI-driven gains go add up to a new pattern of who has money, time, and security — and who doesn’t.

Two Lives in the Same Transition

Seen from a distance, Alex and Rosa are data points in a trend line. Up close, they are living in bodies that have to absorb what the trend line does.

Alex: Teleporting to the summit

Alex does not think of himself as a futurist. He thinks of himself as someone who gets things done.

When the AI tools arrive, he treats them like any other piece of software: something to poke. He is surprised by how quickly they move from curiosity to habit. What begins as a trick for writing cleaner emails turns into an invisible layer across his day: the assistant that drafts, suggests, and reminds.

The first big project he does with heavy AI support is a revelation. He has had the idea for months, but there was never enough time to pursue it. With the new tools, he can generate a plausible outline, refine it, produce visuals, and test variations in days. The internal presentation lands well. He feels, for a moment, like the version of himself he always meant to become.

Then his calendar updates.

People start slotting new work into the space created by the old. The gap between what is technically possible for him and what is emotionally sustainable narrows. He reads faster, writes faster, approves faster. The number of decisions he is expected to make in a week climbs.

If the earlier phase of his career was like hiking, with time to feel the change in elevation, this phase is like being dropped off by helicopter at higher and higher altitudes. He can breathe for now, but there are moments of light-headedness. Somewhere in the back of his mind is a growing worry: Can I actually think this fast about things that matter?

He also notices something subtler. The more polished his outputs become, the less personal credit they seem to attract. In meetings, when someone praises a document, a colleague will occasionally ask, “Did you do this with the AI?” The question is practical, but the implication is familiar: if the invisible exoskeleton did part of the lifting, perhaps the feat is less impressive.

It is a strange emotional equation: more power, less awe. The work is better, but the story about how special it is erodes.

Rosa: Working for the dashboard

For Rosa, the experience is less like teleportation and more like finding that someone is quietly redrawing the map of her trail.

At first, the chatbot shows up at the edges of her job. It handles simple billing questions and status checks. Customers who reach her have already passed through several layers of automation. Some are grateful to talk to a person. Others arrive with the particular frustration of someone who has just argued with a machine.

Then the metrics arrive. Each call is scored by an AI model for adherence to the script, tone, and length. The scores appear in a dashboard that her manager reviews weekly. Feedback shifts from “I heard you rush that customer” to “Your empathy score dipped by 7% last week.”

The logic is not incomprehensible. The company wants consistency. Customers want quick resolutions. Supervisors are managing too many people to listen to every call. But the effect is that Rosa’s sense of being seen by humans dims. A report is looking at her instead.

When a restructuring is announced — “made possible by efficiency gains” — some of her colleagues are laid off. Others are moved into uncertain roles. Rosa’s hours become less predictable. She sees the connection between the system that measures and the decisions that now arrive with less explanation. The fear that had hovered in the background becomes more concrete: The tools that help run my day also make it easier to imagine that someone else could do this, or that fewer someones are needed.

She is not opposed to technology. She uses it all day. What unsettles her is that each new tool seems to increase the distance between her and the people making choices about her life.

If Alex’s story is one of being pulled upward faster than he expected, Rosa’s is one of feeling the ground shifting beneath her feet.

The Emotional Weather of an AI Workplace

Charts can show how many jobs are affected by AI. They cannot show what it feels like to live under those numbers.

For Alex, the emotional climate is sunny with a persistent, high-pressure system:

• Exhilaration. He experiences, directly, the joy of doing in hours what once took weeks. There is a real sense of possibility: side projects that no longer have to be “side.”
• Pride, slightly diluted. He knows that the quality of his decisions and taste still matters. He also knows that some people now assume the hard part is done by the machine.
• Strain. The exoskeleton does not make him superhuman; it simply lets him behave like one for longer than is comfortable. The pace of decisions and the amount of information to absorb now press on his evenings and his sleep.
• Quiet anxiety. He wonders what happens if he stops being able to move this fast. Once you have teleported a few times, expectations do not drift back to the old hiking schedule.

For Rosa, the emotional forecast is different:

• Low-grade fear. Each new automation announcement carries the same unspoken appendix: and so we may need fewer of you. Even when she survives a round of cuts, the feeling remains.
• Distrust. Being measured by systems you cannot inspect makes it harder to believe that feedback is fair. If a human supervisor misjudges you, you can at least talk. How do you talk to a model?
• Loneliness. As more interactions are mediated by dashboards and bots, there are fewer casual, human moments of recognition. The work feels more like a series of transactions and fewer like relationships.
• Resentment. When the same company that touts “AI efficiency” keeps wages flat or shifts people into more precarious schedules, it is difficult not to conclude that the technology is working well — just not for you.

These emotional states are not incidental. They shape behavior. A workforce full of Alex-types who feel both empowered and overextended, and Rosa-types who feel both necessary and precarious, will not respond to change the way a calm, secure workforce does.

They will resist some tools, embrace others, leave suddenly, stay miserably, burn out in place. They will carry their feelings into the voting booth, into conversations with friends, into choices about what to buy and whom to trust.

If AI is a new exoskeleton for the economy, emotional weather is the condition of the body inside it.

Jobs, Incomes, and the Consumer Loop

Zoom out from these two workers, and the pattern becomes systemic.

On the jobs front, AI does not abolish work. It reorganizes it.

• In many fields, task-heavy middle roles shrink. Some are automated outright; others are consolidated into fewer, more demanding positions.
• AI-fluent workers who can design, direct, or meaningfully supervise the tools see their value rise. They are not necessarily engineers. They are people who understand the domain and can articulate clear instructions in language a system can interpret.
• Service and care roles persist, because human bodies and minds still need other human bodies and minds. But these roles increasingly sit under systems that track performance with a precision that previous generations of workers did not experience.

On the income front, the classic question reappears with new technology: who captures the productivity gains?

If firms use AI’s efficiencies to justify wage stagnation, headcount reductions, and concentrated profits, inequality widens even in a growing economy. People like Alex may do reasonably well; people like Rosa see volatility without many paths upward. The middle of the income distribution thins out, not only because jobs disappear but because the quality of the remaining jobs diverges.

This, in turn, touches the consumer economy that underpins much of contemporary life. Every shipping center, streaming service, and grocery chain depends on a large base of people with enough stable income to spend without constant fear. When work becomes more precarious for a significant share of the population, spending patterns shift. People delay big purchases, cut back on discretionary items, and build an economic life around caution.

In one sense, AI could make many things cheaper. Software can design, recommend, and optimize at low marginal cost. But cheaper goods do not compensate indefinitely for erratic or insufficient income. A society where a small group enjoys extremely high productivity-linked earnings and many others live in unstable conditions is one in which certain markets thrive and others wither. It is also one in which collective projects — from public infrastructure to climate adaptation — are harder to fund and harder to agree on.

Finally, there is the social fabric.

Work is not only a way to earn money. It is a source of meaning, routine, and connection. When AI turns some jobs into high-altitude performance feats and others into tightly monitored routines, the opportunities for dignity and belonging stir uneasily.

Workers like Alex may struggle with burnout and a sense that their achievements are being flattened into “what the tools can do.” Workers like Rosa may struggle with feeling unseen and replaceable. Both may feel, in different ways, that they are working inside systems that are optimized for something other than their well-being.

A society full of people who feel this way will not be neutral toward further deployments of AI. It will not simply “adapt.” It will argue, stall, polarize, and occasionally break things.

The Part That’s Still Up to Us

Alex and Rosa did not choose the terms of their encounter with AI. Those terms emerged from a chain of decisions: vendors pitching efficiency, executives choosing metrics, boards approving budgets, policymakers setting or failing to set guardrails.

For an individual worker, the situation can feel like weather — something to be endured, not shaped. But systems are made of choices, and choices can be rethought.

There are at least three places where the story is still pliable.

1. How individuals use and share the tools

People like Alex sit at an interesting fulcrum. They discover, often in private, how powerful these tools can be. They can choose to use AI as a way to get as far ahead as possible, or as an opportunity to make their teams more capable.

This is not a moral test so much as a practical one. When only a few people know how to work with AI, they become bottlenecks and targets. When skills are shared, the workload can be more evenly spread, and the case for supportive, not punitive, use of the tools is easier to make.

People in Rosa’s position have less direct control over the design of systems, but they are not voiceless. They can document what the tools miss, how metrics misrepresent the job, where automation creates new problems, and bring that evidence into conversations with managers, unions, or regulators. The more precise the testimony, the harder it is to dismiss.

2. How organizations define success

Companies face their own fork.

One path is to treat AI primarily as a cost-cutting device: reduce headcount, squeeze more performance out of those who remain, hand as much supervision as possible to dashboards. This can work in the short term. It also tends to produce the kind of emotional weather described earlier: anxiety, distrust, and attrition.

The other path is to treat AI as a tool to reshape work around human strengths: to offload drudgery, allow more time for the parts of the job that require empathy and judgment, and share productivity gains in ways that employees can feel in their pay, their schedules, or their opportunities.

Neither path is pure. Most organizations will mix elements of both. But being explicit about the direction matters. So does including workers in the design of AI deployments, rather than presenting them with finished systems and asking them to adapt.

A company that says, in effect, “Our goal is to keep the exoskeletons as support devices, not as new bodies we hang your job on,” will make different design choices than one that quietly measures success in jobs eliminated.

3. How societies write the rules

Beyond individual workplaces, societies have choices to make about the environment in which these tools operate.

Tax systems can encourage or discourage the hoarding of AI-driven gains at the top. Labor laws can define what kinds of algorithmic management are acceptable, what transparency workers are entitled to, and what recourse they have when an automated system makes a consequential mistake.

Education systems can recognize that learning to work alongside AI is not an optional add-on for a small group but a basic skill of modern life — not in the sense of turning everyone into a programmer, but in the sense of giving more people the fluency that lets them be Alex rather than an anxious bystander.

Social safety nets can cushion the transitions that AI will inevitably accelerate. Not everyone can or wants to teleport to new mountaintops. Periods of retraining, relocation, or simply recovery will be part of the story for many. Whether those periods feel like freefall or like supported movement will depend less on the tools and more on the policies around them.

None of these decisions will fully satisfy everyone. No single law or company policy will transform AI from a source of anxiety into a universal good. But they can bend the curve.

When we talk about AI and the future of work, it is tempting to look for a single answer: utopia or mass unemployment, silver bullet or existential threat. The reality unfolding in offices, warehouses, and call centers is messier.

AI is making some people’s work easier and more interesting, and some people’s work harder and more brittle. It is increasing productivity in ways that strain the minds and bodies of the people who are expected to keep up. It is enabling new services and quietly eroding old forms of security.

Alex and Rosa will not be the last to discover that their lives have been attached to a new kind of exoskeleton. The question, for the rest of us, is whether we are content to let the fit and the load be determined by inertia, or whether we can use the time we have now — before the patterns harden — to ask more deliberate questions.

Not “Will AI take all the jobs?” but “What kinds of jobs, and lives, are we willing to build around it?”