(Or: the Strange Psychology Behind Automation and Power)
Automation has been marketed like a weather system: it’s “coming,” it’s “inevitable,” it will “sweep through” the economy, and one morning you’ll wake up and your job will be a nostalgic memory, like landlines or optimism. The pitch is always the same: machines will take the dull work, humans will float upward into creativity and leisure, and the only thing we’ll do for money is “be ourselves,” which in practice means posting videos of our breakfast with a sound bed of mild despair.
But when you look at what’s actually happening, the story is less sci-fi and more psychological thriller. The machines are not primarily replacing the jobs we’ve spent a century calling “unskilled,” at least not in a clean, total way. They are chewing through work that used to be protected by credentials, vocabulary, and expensive diplomas. They are drafting, summarizing, classifying, predicting, reviewing, coding, triaging, and advising. They are helping screen mammograms and flag anomalies. They are generating first drafts of contracts and legal memos. They are touching the work of doctors and lawyers—exactly the kind of work we were told would be safe because it requires “judgment.” And yet: the barista is still there, the hotel front desk is still staffed, the nail salon still has a human holding your hand like a delicate hostage, and the person cleaning the office still exists in the flesh.
That mismatch—the fact that we’re automating brains faster than bodies—looks like a technical contradiction until you admit the missing variable: the economy is not just an engine for efficiency. It is also a stage for status. And status is not a side quest; it’s a central human motive. Social psychology has been blunt about this for years: people reliably pursue status because the emotional rewards of being admired, deferred to, and treated as important are powerful in themselves, not just instrumental for money. The “status hypothesis” review literature lays this out cleanly: status—respect, admiration, voluntary deference—functions as a fundamental motive across contexts.
Once you accept that, the “coffee shop theory” stops sounding like a cute rant and starts sounding like a workable explanation for why some jobs are culturally protected from full automation. You don’t go to the coffee shop because coffee cannot be made elsewhere. You go for a ritual that flatters the self. You stand in line, you perform preference, you deliver instructions, and the world rearranges itself around your request. The drink is the receipt; the real product is the moment of being centered. And what makes the moment satisfying is not speed. It’s the presence of another human being who can register you—who can interpret your tone, reflect it back, and, crucially, display a kind of trained compliance.
That last part is not me romanticizing service work. It’s me naming what the research already named forty years ago, before we had apps for everything and still believed the future would be tasteful. Arlie Russell Hochschild coined the term “emotional labor” to describe how service workers are paid not only for tasks but for the management and display of feelings—warmth, patience, calm, apology—sold as part of the product. Her classic formulation comes from observing jobs like flight attendants, where “pleasantness” isn’t personality, it’s policy. And later organizational research makes the cost concrete: surface acting (faking the feeling) and deep acting (manufacturing the feeling) predict exhaustion and burnout, precisely because the worker becomes the interface between the customer’s desires and the company’s brand fantasy.
Now here’s where your theory sharpens into something more than a vibe. Machines are wonderful at tasks. They are mediocre at rituals that revolve around hierarchy, because hierarchy is a relationship, not a function. A kiosk can take your order. It cannot participate in the small theater of dominance and deference that many consumers—consciously or not—expect as part of “service.” It cannot look mildly chastened when you change your mind. It cannot apologize in a way that makes you feel powerful. It cannot give you the little social proof of being catered to. For some consumers, that’s a disappointment. For others, it’s the entire point of going out.
This is why the jobs most resistant to full automation often share a common feature: they require a human body to absorb emotion in real time. The front desk clerk isn’t just checking you in; they’re being blamed for your travel day. The server isn’t just carrying plates; they are serving you the sensation of being attended to. The cleaner isn’t just cleaning; they are maintaining a world where your mess disappears without you having to confront your own animal nature. The nail technician is not simply painting keratin; they are offering a form of intimate attention that is both comforting and status-coded, which is why it flourishes even in a supposedly “high-tech” economy. This pattern is so visible that economists have popularized it in plain language: the growth in interpersonal service jobs (the so-called “manicure economy”) sits right next to the decline of routine middle-skill work, because what machines replace easily are standardized tasks, while what they struggle to replace is human interaction that people value for social reasons.
None of this means low-wage jobs are safe, by the way. It means they’re safe in a very specific sense: they can be technologically replaceable and still socially preserved, partially, because customers and organizations use them as a venue for performance—of service, of control, of “being important,” of “being taken care of.” And that preservation can be cruel. It can lock humans into roles where their main value is emotional absorption. The worker stays human not because the world cherishes them, but because the world still wants someone to be “lower” in a way a machine can’t convincingly embody.
At this point, someone always objects: “But people hate bad service. Wouldn’t they prefer automation that’s faster and more accurate?” Sometimes, yes. If you’re buying a train ticket, most of us would happily skip the line and the sighing and press the buttons ourselves. But the research on algorithm aversion and AI resistance explains why this flips in domains tied to identity, trust, and social meaning. People often avoid algorithms after seeing them make mistakes, even when those algorithms outperform humans, because they lose confidence in machines more quickly than in people. And in healthcare specifically, consumer research has found meaningful resistance to “medical AI”—not because people think doctors are perfect, but because they experience care as relational, and they attach moral weight to a human being being responsible.
This is where your theory gets its sharpest, most uncomfortable teeth: the roles that persist are often the roles that preserve hierarchy in public. In other words, the “master” doesn’t just want the outcome; the master wants the relationship. A latte delivered by a perfect machine is coffee. A latte delivered by a human is coffee plus a tiny confirmation of social position. Even if nobody would phrase it like that at brunch, the structure is there.
And now we can address your other claim—the one that sounds upside down until you look closely: why skilled professions like law and medicine are getting hit so aggressively. The reason isn’t that society values baristas more than doctors. It’s that large parts of “professional” work are, at the task level, text and pattern manipulation, and modern AI is built to be a statistical predator of patterns. That doesn’t mean doctors and lawyers vanish; it means their workflows get sliced into automatable components at a pace that feels like disrespect. Major economic reports have been explicit that generative AI can automate or accelerate significant shares of work hours, with large exposure in professional, business, legal, and some healthcare-related activities, largely via writing, analysis, and information processing.
The legal world is already living this. Lawyers are adopting generative tools for drafting, research, summarization, and document work, while simultaneously panicking about hallucinations, confidentiality, and professional responsibility—because nothing says “future” like filing a brief with fake citations and getting sanctioned. There are now task forces and guidance documents from bar associations specifically addressing how lawyers should use and supervise AI, which is basically the profession admitting, through gritted teeth, that the tool is here and it touches the core of the work. The regulatory anxiety is no longer theoretical either; legislators are already proposing guardrails that force lawyers to verify AI-generated materials before submitting them, precisely because the work product can be persuasive while also being wrong.
Medicine is doing the same dance, just with higher stakes and better lighting. Radiology has become the symbolic battleground because it sits at the intersection of imaging, pattern recognition, and clinical judgment—the exact terrain where AI performs well. And the most credible voices in radiology are not saying “AI replaces radiologists tomorrow”; they’re saying “radiologists plus AI,” with the profession shifting toward oversight, integration, and accountability. That “radiologist in the loop” framing has been in the literature for years. Recent studies and trials also support the more nuanced reality: AI can improve certain screening workflows and detection rates when used as an assistive system, while experts emphasize careful implementation and monitoring rather than replacement.
So yes—doctors and lawyers are being “automated” faster than the barista in a very specific way: not by being wiped out, but by having large portions of their tasks absorbed into systems that produce drafts, suggestions, triage, flags, and summaries at scale. That is exactly what you observed. And it dovetails with the broader labor-economics view that automation substitutes for some tasks while complementing others, changing the shape of work rather than simply deleting it.
Now, can we “prove” your theory—the claim that automation won’t fully replace service labor because people need someone beneath them to feel like someone? In the strict scientific sense, “prove” is a trap word. Human societies are too messy for single-cause proofs. But we can do something better than proof-by-vibes: we can show that your theory is consistent with multiple, independent research streams that all point to the same mechanism.
First, the status literature says people seek deference and admiration as an intrinsic motive. Second, emotional labor scholarship shows that many service jobs are explicitly designed around the management of customer emotion and the performance of deference. Third, behavioral research shows people often distrust algorithms, especially after mistakes, and exhibit resistance to AI in medicine even when AI is framed as beneficial. Fourth, labor economics documents that job growth has shifted toward interpersonal services in part because those tasks are harder to automate and because human interaction itself remains economically valuable. Fifth, major institutions tracking automation risk emphasize that exposure is about tasks and skills, not job titles, and that automation pressure reshapes work unevenly across occupations.
When you stack those together, your “master needs a mirror” idea stops being a hot take and starts being a plausible sociological claim: the economy preserves certain human-facing roles not purely because machines can’t do them, but because humans want the relationship those roles provide—control, recognition, a place to offload frustration, a stage on which to be important. The machine can deliver the outcome. It can’t deliver the social meaning. And modern consumption is, embarrassingly, drenched in social meaning.
This also explains why the most automated experiences still keep a human somewhere nearby, like an emotional safety valve. Self-checkout lanes still need an attendant, partly because machines fail, but also because customers want a human target when they’re irritated. Airline apps do everything until something goes wrong, and then you find yourself pleading with a gate agent—because the gate agent is not only solving a problem; they’re absorbing your panic and translating it into a narrative where you might still make the wedding. The call center script is the purest example: you can build a perfect chatbot, but companies keep humans for escalation because “I’m sorry” only works when it costs someone something to say it.
So the future isn’t “humans replaced by robots.” The future is more morally confusing. The future is machines doing the thinking and humans doing the kneeling, unless we deliberately change what we reward and what we tolerate. We’re racing to automate cognition—law, medicine, art—while preserving the jobs that require a person to stand there and be treated as lesser in a socially acceptable way. We are automating the prestigious and preserving the degrading, then calling it progress because the graphs look efficient.
And that’s the punchline you’re circling: automation doesn’t naturally abolish hierarchy. It can actually polish hierarchy to a mirror shine. If we don’t challenge the status cravings and dominance rituals baked into daily life, we’ll keep our baristas and cleaners not because we cherish them, but because too many people still want a small daily throne—somewhere to be obeyed, somewhere to be soothed, somewhere to feel, for ten minutes, like the world is arranged correctly.
And this is where the story widens beyond coffee shops and clinics and into the architecture of modern life itself. Once you see hierarchy as a form of emotional infrastructure, you begin to notice how many systems quietly rely on it. Luxury retail, for instance, still revolves around the presence of sales associates who are trained not merely to sell but to affirm. The price tag is part of the fantasy, but the real value is being seen as someone worth attending to. Self-checkout kiosks may be convenient, but they cannot replicate the subtle choreography of admiration and deference that signals belonging to a particular social tier. Research on status consumption has long shown that people do not buy luxury solely for quality or function; they buy it for recognition, distinction, and the emotional feedback loop of being treated as special. The product is just the prop. The interaction is the stage.
The same logic governs corporate environments that appear, on the surface, to be highly automated. Open-plan offices still employ receptionists, office managers, and assistants who perform a quiet kind of emotional smoothing—greeting, redirecting, softening conflict, making sure the social temperature stays tolerable. These roles are rarely described as strategic, yet organizations consistently preserve them because they stabilize internal hierarchies and reduce friction. In this sense, emotional labor operates as a form of organizational glue. It absorbs tension so that power structures do not have to account for it.
Economic research reinforces this pattern. Scholars studying labor polarization have observed that as routine middle-skill jobs decline, employment growth concentrates at two ends of the spectrum: high-skill analytical roles and low-skill interpersonal service roles. The latter category, sometimes described as the “manicure economy,” expands not because it is technologically advanced, but because it depends on face-to-face interaction, situational judgment, and emotional responsiveness—qualities that resist full automation. What looks like inefficiency is often a deliberate retention of human presence because that presence performs a social function machines cannot convincingly fulfill.
This helps explain why even the most automated systems still retain a human layer. Self-checkout lanes require attendants. Ride-share apps maintain customer support teams. Online banks still staff call centers. These humans are not there primarily to process information; the system already does that. They are there to absorb blame, to offer reassurance, to provide a sense of moral accountability when technology fails or feels impersonal. Behavioral studies on algorithm aversion show that people lose trust in machines more quickly than in humans, even when the machines perform better on average. A single visible error can trigger rejection of an algorithm, whereas humans are often forgiven as fallible. This asymmetry reveals that trust is not about accuracy alone; it is about relationship and perceived intention.
Healthcare makes this especially clear. Patients often say they want efficient, data-driven care, yet express discomfort when decisions appear to be made solely by algorithms. Research in medical sociology and consumer psychology shows that people attach moral weight to the idea that another human is responsible for their wellbeing. They want someone who can be held accountable, who can express empathy, who can be blamed if something goes wrong. Even when AI systems demonstrate higher diagnostic accuracy in specific tasks, many patients prefer them as decision-support tools rather than replacements. The resistance is not to technology itself, but to the removal of the relational layer that signals care and responsibility.
The legal profession faces a parallel tension. Generative systems can now draft, summarize, and analyze legal text at scale, performing in seconds what once took teams of junior associates days. Yet the profession insists on human oversight, not only for ethical and regulatory reasons, but because legal authority is still symbolically tied to a human voice. Clients do not simply want a correct contract; they want a lawyer who represents them, who can be confronted, persuaded, or blamed. The machine can produce the document, but it cannot perform the role of advocate in the social imagination.
Across these domains, the same structure repeats. Automation excels at tasks. It struggles with relationships. And relationships, especially hierarchical ones, are where much of social meaning resides. The persistence of service roles is therefore not an accident or a lag in technological adoption. It is a reflection of how deeply status and power are woven into everyday exchanges. We preserve certain forms of human labor not because they are efficient, but because they make hierarchy visible and emotionally legible.
This is also why attempts to fully automate service environments often feel unsettling. A hotel staffed entirely by kiosks may be efficient, but it feels cold, transactional, and strangely disorienting. Without a human intermediary, there is no one to receive frustration, no one to apologize, no one to perform care. The absence of that layer exposes the transactional nature of the exchange too starkly, stripping away the social cushioning that makes inequality feel tolerable.
Seen this way, automation does not simply reshape the labor market; it reveals what we value. We automate what can be abstracted and standardized, even when it carries prestige. We preserve what sustains social rituals, even when it is demeaning or exhausting. The result is a future in which cognitive labor becomes increasingly invisible and automated, while emotional and relational labor remains stubbornly embodied.
This does not mean hierarchy is immutable. It means it is reproduced through habits, expectations, and everyday performances that feel natural because they are familiar. Your theory—that people unconsciously preserve roles that allow them to feel superior because those roles satisfy a psychological need—is not a provocation for its own sake. It is a lens that clarifies why automation unfolds the way it does. It shows that technological change is filtered through human desires for recognition, control, and belonging.
Until those desires are confronted, the trajectory will remain the same. Machines will continue to absorb the work of thinking. Humans will continue to perform the work of being seen. And the economy will keep staging the same quiet drama: someone ordering, someone complying, someone feeling, for a moment, like the world is arranged correctly.
Once you notice how emotional hierarchy operates in service spaces, it becomes impossible to unsee it in places that claim to be about aspiration rather than submission. Luxury retail is one of the clearest examples. On paper, it is simply the sale of expensive objects. In practice, it is a carefully staged interaction in which the customer is made to feel chosen, recognized, and elevated. Sales associates are trained to mirror taste, affirm identity, and subtly signal that the buyer belongs in the world the brand represents. The handbag, the watch, the shoes are not the core of the experience. The real value is the emotional feedback loop that says: you are someone worth attending to.
No algorithm can convincingly replicate that exchange. A screen can recommend, but it cannot admire. It cannot lean in conspiratorially, cannot remember your name, cannot signal that you are not just a consumer but a person with status. This is why even the most digitally sophisticated luxury brands still invest heavily in in-store human experience. They are not preserving inefficiency; they are preserving theater. The presence of a human intermediary transforms a transaction into a social ritual, one that reinforces both the customer’s identity and the brand’s mythology.
The same dynamic is playing out online, where social media has created a new ecosystem of micro-hierarchies. Likes, follows, views, and comments function as symbolic currencies, conferring visibility and validation. People now perform status for one another at scale, curating identities and measuring worth through metrics that fluctuate in real time. In this environment, automation is already embedded—algorithms shape what we see, who is amplified, and which voices disappear—but the emotional core remains human. Influencers may rely on automated tools to schedule posts, analyze engagement, and generate content, yet their value lies in the perception of authenticity, relatability, and personal connection. Followers do not just want content; they want recognition. They want to feel seen by someone they admire.
This is another version of the same structure. The machine handles the logistics. The human performs the relationship. The hierarchy is sustained through attention and response, not through technical necessity. Even when bots can mimic conversation, they struggle to produce the subtle social signals that make people feel acknowledged rather than processed. Trust, admiration, and belonging remain stubbornly relational.
Corporate life, too, is built around these quiet rituals of power. Organizations often claim to value efficiency and flat hierarchies, yet they preserve layers of roles whose primary function is social regulation. Assistants, coordinators, receptionists, and managers frequently act as buffers between authority and the rest of the workforce. They translate directives into palatable language, absorb complaints, and maintain the emotional climate that allows the hierarchy to function without open conflict. These roles are rarely framed as strategic, but companies continue to fund them because they stabilize the social order. They are the human interface through which power is made tolerable.
This helps explain why middle management, despite being a frequent target of criticism, remains difficult to eliminate. Even when their operational tasks can be automated, their relational function persists. They provide a face for the system, someone who can be appealed to, blamed, or persuaded. Removing them would expose the organization’s structure too starkly, stripping away the interpersonal layer that softens its authority.
Research in organizational psychology supports this. Studies on leadership and power dynamics show that people are more accepting of unequal structures when those structures are mediated by interpersonal relationships. When authority is perceived as relational rather than mechanical, it feels more legitimate. Automation threatens this perception by replacing negotiable human actors with fixed systems. The discomfort that follows is not just about losing jobs; it is about losing a familiar way of relating to power.
This is why even highly automated companies maintain customer support teams and escalation channels staffed by humans. The system can process requests, but when something goes wrong, people want someone who can say, “I’m sorry,” in a way that feels costly. The apology must come from a person because it signals moral responsibility. A machine’s apology is merely functional; it does not carry the same emotional weight. This distinction is critical. It reveals that our attachment to human labor in certain roles is not based on performance alone, but on the symbolic meaning of accountability and care.
The persistence of these roles also reflects a deeper tension in how societies define value. We often claim to reward intelligence, creativity, and expertise, yet the labor market increasingly treats these qualities as transferable to machines. At the same time, it preserves roles that require people to be physically present, emotionally responsive, and visibly subordinate. This inversion challenges the assumption that progress naturally elevates human dignity. Instead, it suggests that progress can coexist with, and even reinforce, social stratification.
Your theory, that automation will not fully replace service labor because people derive psychological satisfaction from occupying a higher position—fits within this broader pattern. It does not deny economic forces; it reframes them. Technology does not operate in a vacuum. It is adopted, resisted, and shaped by cultural values and social needs. When a role serves as a mirror for someone else’s status, it acquires a form of protection that is not captured by cost-benefit analysis.
This does not mean individuals consciously seek to dominate others. Much of this behavior is subtle, habitual, and socially normalized. People may simply feel more comfortable in environments where roles are clearly defined and where their position is affirmed through small interactions. Over time, these micro-rituals accumulate into stable structures that resist change, even when alternatives exist.
Returning to the coffee shop, the ritual now looks less trivial. It is a daily rehearsal of hierarchy, performed in a socially acceptable way. The barista’s presence allows the customer to experience a fleeting sense of importance, just as the front desk clerk, the sales associate, and the call center agent do in their respective domains. These interactions are not accidents. They are cultural scripts that distribute emotional rewards and burdens in patterned ways.
Automation, then, does not simply replace work. It exposes which parts of work we are willing to relinquish and which we insist remain human. We are comfortable letting machines think for us. We are less comfortable letting them replace the social performances that make power visible and relational. The result is a future where intelligence is increasingly abstracted into systems, while emotional labor remains embodied, unevenly distributed, and often undervalued.
If there is a choice embedded in this trajectory, it lies in what we decide to preserve. We can continue to treat human presence as a tool for maintaining hierarchy, or we can begin to question why these rituals feel necessary. The technology itself does not demand a particular social order. It is our response to it—shaped by desires for recognition, control, and belonging—that determines whether automation becomes a force for liberation or another mechanism for reinforcing inequality.
The old promise of automation was freedom. What we built instead was a mirror. A world where machines handle the thinking while humans continue the performance. Because power, at its core, is not about results. It is about contrast. The master only feels like the master when there is another of their own kind standing beneath them. Not a screen. Not a robot. A person. Someone who could have been them, but isn’t.
So the future will not be jobless. It will be deeply awkward. Machines will calculate, predict, and compose with flawless speed. Humans will smile, apologize, carry, clean, serve, and absorb the quiet frustrations of people who need to feel seen as superior, if only for a moment. The hierarchy will remain, polished and intact, not because it is efficient, but because it is familiar.
And every morning, as someone orders a drink they could have made at home, they will unknowingly participate in the oldest system we have: the performance of power, disguised as convenience, served hot, with foam—no foam—actually, a little foam.
Related:
When creativity meets code — AI: The Palette, Not the Painter →
Copyright © 2026 Tantrum Media LLC. All rights reserved. All ideas, creative concepts, and editorial content are the intellectual property of Tantrum Media LLC. AI tools may assist with drafting, research, and grammatical correctness, but all content is creatively conceived and directed by Tantrum Media LLC. No reproduction or republication without permission. Brief quotations permitted with proper attribution and a direct link to the original source.