HEADSHOTLAB
The Lighting Was Never the Point: On the Strange New Honesty of AI Generated Headshots
There is a particular silence that falls over a photo studio in the half-second before the shutter clicks. The assistant has just said "beautiful," the softbox hums faintly, and the subject—usually a lawyer, a founder, a mid-career marketer rebranding herself for a new chapter—holds an expression that is not quite a smile and not quite anything else. It is the face of someone trying to look like the best available version of their own LinkedIn photo. That silence, and the four hundred dollars and ninety minutes it usually costs to produce it, has become strangely optional.
Somewhere in the last three years, without much fanfare, the professional headshot quietly detached itself from the studio altogether. It now happens, for a growing number of people, on a laptop at midnight, in pajamas, using a folder of selfies dragged from a phone. The output looks, by most measures, indistinguishable from the studio version: the blazer is pressed, the background is a tasteful gradient of charcoal or sage, the eyes catch a light that was never actually in the room. This is the world of the AI generated headshot, and it has arrived with the kind of unglamorous suddenness that tends to precede a genuine shift in how people present themselves to the world.
The Portrait as Performance, Reconsidered
Portraiture has always been a negotiation between the sitter and the circumstances. The Flemish burghers who commissioned oil paintings of themselves in the sixteenth century were not interested in documentary accuracy; they wanted gravitas, fine fabric, the suggestion of a library behind them even if they owned four books. The corporate headshot inherited this instinct almost wholesale. It is not meant to capture a person as they appear on an ordinary Tuesday—it is meant to capture them as they'd like to be filed away in a stranger's memory: competent, approachable, faintly successful.
What makes the AI generated headshot culturally interesting is not that it automates this fiction, but that it makes the fiction explicit. Nobody uploading selfies to a service like Headshot Lab is under any illusion that the resulting portrait is a literal photograph. It is understood, in the way a retouched magazine cover is understood, as a composite—part likeness, part aspiration, rendered by a model trained on the visual grammar of corporate photography: the three-point lighting, the slight downward gaze into the lens, the blazer that has never touched a dry cleaner because it does not, in the traditional sense, exist.
This is a departure from decades of portrait photography etiquette, in which retouching was meant to be invisible—present but denied, like stage makeup under television lights. The AI headshot inverts that discretion. Its artificiality is the point, openly negotiated between user and interface: choose the backdrop, choose the attire, choose how many teeth to show. Susan Sontag, writing in On Photography, warned that photographs "furnish evidence," that a picture passes for incontrovertible proof a thing happened. The AI headshot quietly breaks that contract and nobody seems especially bothered. The image is not evidence of a photo shoot. It is evidence of effort—of someone caring enough about their professional presentation to curate it, even if the curation happened inside a neural network rather than a camera bag.
The Economics of Looking Employable
Ask any career coach what the single most consequential image in a person's professional life is, and most will say the same thing: it's the headshot, not the résumé. LinkedIn's own research has repeatedly found that profiles with a photo receive dramatically more views and connection requests than those without, and recruiters routinely admit, off the record, that an unprofessional or absent photo primes an unconscious bias before a single line of experience is read. The headshot, in other words, is not decoration. It is infrastructure.
And yet for most of photographic history, a good headshot was a function of disposable income and geography. You needed a photographer willing to travel to you or an office with a space to convert into a makeshift set; you needed hair and makeup, or at least the confidence to fake it; you needed, crucially, a free weekday afternoon and somewhere between two hundred and a thousand dollars depending on your city. In London or New York, a half-day corporate shoot with retouching could easily run past a thousand pounds. For a freelancer in a smaller market, or a graduate student applying for their first postdoctoral position, that cost was not an inconvenience—it was prohibitive.
The arrival of services that generate dozens of polished headshot options from a handful of uploaded selfies collapses that economics almost entirely. A single fee, often less than the cost of one hour with a traditional photographer, now returns a gallery of looks: different backgrounds, different attire, different expressions, enough variety that a user can select one for their company's website, a slightly warmer one for a conference bio, a more severe one for a legal directory. Headshot Lab built its offering around precisely this asymmetry—turning a handful of phone photos into a wardrobe of professional options, with retouch credits for the inevitable moment someone notices their collar sits oddly, and a money-back guarantee for the rarer moment the whole batch simply misses the mark.
This matters most, perhaps, for the people the traditional headshot economy quietly excluded: the remote worker who has never set foot in their company's office and has no local studio recommendation, the job seeker between paychecks who cannot justify spending on a photo before securing the position it's meant to help win, the small nonprofit that needs twelve consistent staff portraits and has a marketing budget measured in hundreds, not thousands. For these people, the AI headshot isn't a novelty. It's a door that was previously locked.
What the Machine Learned From Watching Us Perform
There is something almost anthropological in asking what, exactly, these systems were trained to understand about professionalism. The models behind contemporary headshot generators have absorbed an enormous visual corpus of what "looking competent" has historically meant in Western corporate culture: symmetrical lighting, neutral or gently blurred backgrounds, muted color palettes in clothing, a certain restraint in expression that reads as confidence rather than anxiety. In doing so, they've effectively codified decades of implicit workplace aesthetics—the unspoken rules that told women to avoid overly bright lipstick in corporate photos, that told men a crisp collar signals more seniority than an open one, that a navy blazer reads as more trustworthy than black.
This is worth sitting with, because it means the AI headshot is not a neutral technology quietly replicating photography. It is, in a sense, a mirror held up to a very specific and historically narrow idea of what professional belonging looks like—one that originated in mid-century American corporate culture and has since been exported, with minor variations, to boardrooms and recruitment platforms worldwide. Researchers studying algorithmic bias, including those cited in MIT Technology Review's ongoing coverage of generative AI, have pointed out that image models can struggle to render certain skin tones, hair textures, or religious garments with the same fidelity and flattery as others, a flaw inherited directly from imbalanced training data. The better services in this space have had to work deliberately against that inheritance, widening their datasets and offering more granular style controls so that a hijab, locs, or a turban are rendered with the same care as a standard blazer-and-tie combination, rather than treated as an edge case.
The cultural stakes here are higher than they might first appear. A headshot is often the first and sometimes only visual information a stranger receives about a person before forming a judgment about their competence, trustworthiness, and seniority—a phenomenon social psychologists call "thin-slicing," a concept popularized by Malcolm Gladwell in Blink and studied extensively in the decades since. If the machine generating that first impression has been trained predominantly on one demographic's version of professionalism, it risks quietly reproducing the exclusions of the old boys' club it was meant to democratize. The companies doing this well are the ones treating representational range not as a feature update but as the whole point.
The Office That Looks the Same From Every Angle
Perhaps the most organizationally interesting use of this technology has nothing to do with individual vanity at all. It has to do with the particular visual chaos of the modern distributed company. Walk through the "team" page of almost any fast-growing startup founded before 2020, and you'll likely find a patchwork of headshots taken at wildly different moments: one employee photographed at an all-hands meeting three years ago under fluorescent light, another who submitted a vacation photo cropped awkwardly at the shoulders, a new hire whose square avatar is still the generic gray silhouette because nobody has gotten around to scheduling their shoot. It is a small thing, but it reads, cumulatively, as disorganization—a visual tell that the company has grown faster than its systems.
Traditional solutions to this involved flying a photographer to a head office once or twice a year and hoping remote staff could make the trip, or sending a local photographer to each satellite office at considerable coordination cost. Team headshot packages built around AI generation solve the problem differently: every employee uploads their own selfies from wherever they happen to be, and the output is unified by background, lighting style, and dress code, regardless of whether the person is in Lisbon, Manila, or a spare bedroom in Ohio. Headshot Lab's team offering reflects this exact use case—organizations wanting the visual coherence of a single studio day without requiring a single employee to leave their desk.
There's a quiet dignity in this that's easy to miss. Visual consistency across a team page is not vanity; it's a form of institutional care, a signal to clients and recruits alike that the organization sees its people as part of one coherent story rather than a loose collection of individuals who happened to end up on the same payroll. Design consultancies have long understood this instinct—Pentagram, among others, has written about brand coherence as a function of trust—and it turns out that principle applies as much to the faces on an "About Us" page as it does to a logo or a typeface.
The Uncanny Valley Gets a Mortgage
None of this is to suggest the technology is without its frictions, and the discourse around AI headshots has, since roughly 2023, developed its own small folklore of cautionary tales: the executive whose AI-generated portrait gave him an extra finger resting on his chin, the job seeker whose teeth were rendered just slightly too white and too uniform, triggering that specific, hard-to-name unease psychologists still refer to as the uncanny valley, a term coined by robotics professor Masahiro Mori and explored at length in outlets like Scientific American. Early generations of the technology were, frankly, more prone to these glitches—smoothed-over skin that read as plastic, backgrounds that dissolved at the edges into an unconvincing blur, an occasional and unsettling sense that the eyes weren't quite looking anywhere.
The technology has matured considerably since then, and the better platforms now return results that colleagues and hiring managers genuinely cannot distinguish from a traditional studio portrait, particularly once a subject has selected from a wide enough batch to find the one or two frames where the model's interpretation of their face lands convincingly. Part of the ongoing improvement has come from more sophisticated handling of texture and asymmetry—real faces are not perfectly symmetrical, and the systems that succeed now deliberately preserve small natural imperfections rather than ironing them out.
There remains, nonetheless, a live ethical question worth naming honestly rather than glossing over: at what point does an idealized AI portrait become a kind of quiet misrepresentation, if the person who shows up to the actual interview or client meeting looks meaningfully different from their photo? The honest answer is that this tension predates AI by a long way—heavily retouched studio photography raised the identical question for decades, and anyone who has matched a dating profile photo to the person who arrives at the café already understands the gap between curated and actual. The responsible position, and the one the more credible services in this space have adopted, is to aim for an enhanced but recognizable likeness rather than a different face altogether: better lighting and tidier styling, not a new jawline. The goal is the best Tuesday, rendered clearly—not a Tuesday that never happened.
A Face for the Algorithm, A Face for the Room
What's happening with headshots is, in miniature, what's happening with professional self-presentation more broadly: the image we show an algorithm—a recruiting platform's search results, a LinkedIn feed's ranking logic, a conference website's speaker grid—now matters as much as the image we show a person across a table. We are, increasingly, curating two faces: the one assessed by other humans in real time, and the one indexed, compared, and surfaced by systems that never blink. The AI generated headshot belongs unmistakably to that second category, and its rapid adoption suggests most professionals have made peace with that duality far faster than the cultural commentary around it has caught up.
There is something almost reassuring, in the end, about how undramatic this shift has turned out to be. The headshot was never really about the studio, the softbox, or the assistant murmuring encouragement before the shutter clicked—it was always about producing a credible, professional version of a face for strangers to trust before they'd had the chance to meet the person behind it. That version now arrives, for many, through a browser tab rather than an appointment, selected from a gallery rather than a contact sheet, finished in minutes rather than a fortnight of waiting for proofs. Somewhere in a server farm, a model is quietly generating thousands of blazers that have never been worn, under lighting that was never switched on, for people getting ready to look, at last, like themselves on a very good day—which is, perhaps, all a headshot was ever meant to do.
Article written by FrontRank