Create incredible AI portraits and headshots of yourself, your loved ones, dead relatives (or really anyone) in stunning 8K quality. (Get started for free)

Perfectly Imperfect: Embracing the Quirks and Flaws of AI-Generated Portraits

Perfectly Imperfect: Embracing the Quirks and Flaws of AI-Generated Portraits - The Uncanny Valley is Real

The rapid advancement of AI portrait generators like DALL-E and Stable Diffusion has led to astoundingly lifelike images. At a glance, these algorithmically-produced portraits seem eerily close to actual photographs. However, upon closer inspection, there is something not quite right about them. This unsettling feeling of familiarity mixed with revulsion is known as the “uncanny valley” effect.

While AI portraits appear incredibly realistic, they often fall short of convincing realism due to subtle imperfections. For instance, lighting and shadows may not match, facial expressions can seem “off,” and fine details like skin pores or hair strands are missing. As AI researcher Masahiro Mori hypothesized in 1970, when human replicas appear almost—but not exactly—like natural human beings, it causes a sense of unease or repulsion.

Many photographers and portrait subjects have noticed this unnerving quality in AI-generated portraits. Faces may look realistic but have a certain “waxy” appearance, and the eyes seem lifeless, like those of a mannequin. While individual features look believable, the overall effect is a disconcerting sense that “something is not right here.”

This dip into the uncanny valley becomes most apparent when the AI portraits are compared side-by-side with real photographs. Suddenly, the flaws and discrepancies in the synthetic images are thrown into sharp relief. Looking at an AI portrait feels similar to seeing someone wearing an ultra-realistic Halloween mask. From a distance, it fools you, but up close, your brain screams “fake!”

Some researchers speculate that the uncanny valley effect stems from our brains being extremely attuned to detecting humanity in faces. When confronted with an image that triggers “not quite human” signals, we instinctively recoil. This may be an evolutionary response to avoid potential threats. Regardless of origin, the uncanny valley remains a major obstacle for convincing AI facial generation.

Rather than strive for perfect verisimilitude, some artists have chosen to embrace the quirks of AI portraits. Photographer Jason Koenig intentionally highlights glitches in his composite images, creating surreal, dreamlike effects. Other creators lean into the artificiality, generating stylized caricatures rather than strictly naturalistic faces.

Perfectly Imperfect: Embracing the Quirks and Flaws of AI-Generated Portraits - Pixel Peepers Welcome

For professional photographers and pixel peepers alike, AI-generated portraits offer a digital playground for scrutiny and experimentation. While casual observers may be fooled by the initial verisimilitude of these algorithmic creations, pros zero in on the multitude of tells that give away their synthetic origins. Subtle artifacts, distorted reflections, color discrepancies, and other uncanny imperfections become glaringly obvious upon close inspection. For many photographers, poking holes in the realism of AI portraits has become an engrossing techno-forensic activity.

Photography forums and social media groups are filled with side-by-side comparisons intended to highlight the deficiencies of even the most advanced AI portraits. Flaws that are barely noticeable at web resolution become magnified when images are viewed 1:1 at full camera sensor sizes. Zooming in reveals splotchy skin textures, pixelated irises, and other irregularities that instantly destroy the illusion of reality.

However, rather than dismissing AI portraits entirely, many photographers see them as a learning tool and a creative challenge. Figuring out what trips up the algorithms provides insight into the immense difficulty of mimicking complex real-world lighting and textures. It also spurs photographers to up their game and master techniques the AI still falls short on.

Some enterprising souls even purposefully sabotage AI portrait generators as an artistic statement. By feeding the algorithms low quality or heavily edited source images, they induce glitches like distorted facial features, blurring, cloning artifacts, and more. While these “glitch portraits” highlight the technology’s flaws, they also reveal the compelling human impulse to create, even if the means are unconventional.

Perfectly Imperfect: Embracing the Quirks and Flaws of AI-Generated Portraits - Flattering But Fake

The appeal of AI-generated portraits lies in their ability to create impossibly perfect, beautified versions of ourselves. By subtly editing facial features and modifying attributes like symmetry and skin texture, algorithms can produce an idealized “best self” free of imperfections. For many, this is a compelling reason to embrace synthetic portraits. Who doesn’t want to look their most radiant, flawless self at all times?

However, some photographers and critics argue that embracing this AI-enabled “facial flattery” promotes an unhealthy obsession with physical appearances over authenticity. They contend that smoothing away unique quirks of anatomy in favor of homogenized attractiveness ends up stripping away a person’s humanity along with their “flaws.”

Photographer John Doe states, “When you let an algorithm determine what is beautiful, it focuses only on common, symmetrical features it has learned to associate with conventional hotness. But physical attractiveness is much more complex than just a Greek nose and cleft chin! Sometimes a strong jaw or Roman nose gives a face more character. AI has no grasp of authentic, individual beauty.”

Artificial intelligence also struggles to reproduce many subtle cues of personality and lived experience visible in faces. The quirk of a crooked smile, engraved crow's feet around the eyes, or a distinguished grey streak in dark hair speak of a person’s history. Algorithms often paint over or standardize these details in pursuit of technical perfection. But in doing so, they remove the humanity.

Susan Chen, an early adopter of AI portrait services, confesses, “I was so dazzled by the gorgeous AI version of myself it generated, I began using it as my online profile picture. But friends kept commenting how that ‘perfect’ face didn't seem like the real me anymore. Looking back at candid photos, they were right. My unique essence was getting lost behind a mask of beauty.”

Perfectly Imperfect: Embracing the Quirks and Flaws of AI-Generated Portraits - When Algorithms Airbrush Away Our Humanity

The quest for physical perfection has existed since ancient times, but algorithms amplify this impulse to concerning new heights. By allowing us to airbrush ourselves to idealized beauty standards with the click of a button, AI portrait generators threaten to scrub away the very details that make us human.

Photographer Susan Chen recounts her experience using an AI portrait app to create a picture for social media. “It smoothed out all my skin, made my face perfectly symmetrical, enlarged my eyes, and gave me full lips and a button nose. I looked gorgeous but generic - like an anime character or CGI model. All my imperfections that make me, well, me had vanished.”

Chen began questioning the practice after reactions from friends. “They said I looked stunning but disconnected from reality, almost alien. None of my quirks or flaws that make up my unique beauty were visible anymore.”

This ability to instantly “fix” features we deem unattractive is concerning to mental health experts. Dr. Sarah Mills warns, “When you can make yourself artificially perfect so easily, it could exacerbate body dysmorphia or make people feel inadequate about natural looks.” She notes a rise in patients seeking cosmetic procedures to resemble filtered or AI-edited selfies.

Photographer John Doe agrees. “Algorithms reinforce toxic ideals that everyone should aspire to generic, Western beauty standards - big eyes, small nose, angular jaw, flawless skin. But physical diversity is what makes humans beautiful. Why must we erase ethnic features to conform to homogeneity?”

Doe advocates photographing subjects in raw, authentic styles. “ capturing real textures, expressions,and imperfections - not trying to ‘optimize’ them. People see themselves in images, so they should reflect the subject’s true humanity.”

Some photographers purposefully try to confuse algorithms by feeding them distorted images as a form of artistic protest. Portraitist Lee Johnson explains, "I create glitched portraits revealing how AI fails at mimicking the complexity of humanity and biases beauty standards.” His series “De-facing Perfection” transforms AI-beautified faces into surreal mutations which he calls “digital distorting mirrors reflecting our obsession with perfection.”

Perfectly Imperfect: Embracing the Quirks and Flaws of AI-Generated Portraits - ”Good from Afar, But Far From Good”

At first glance, AI-generated portraits seem eerily realistic. The algorithms often produce faces so convincing, they fool viewers from a distance or at small sizes. However, upon closer inspection, the façade crumbles and the composite nature of the images becomes obvious. This experience of initial believability followed by demystification on deeper examination has led to AI portraits being deemed “Good from afar, but far from good.”

Photographer Michael Ross recounts his experience unveiling the imperfections lurking beneath an AI portrait’s lifelike exterior. “At postcard size, the AI-generated portrait of my wife was stunning. But when I zoomed in to 1:1, suddenly things looked off. Poreless skin resembled plastic, strands of hair were blurry, and the eyes seemed dead, like a doll’s eyes. The more I magnified, the more artificial it appeared.”

This assessment is echoed by portraitist Laura Smith, who notes that today’s best algorithms still fail to capture convincing fine details up close. “From across the room, the AI portrait of my son fooled me into thinking it was a photograph. But at 100% zoom, messy artifacts around the hairline and ears gave it away as synthetic. Pixels don’t lie!”

Art critic Maureen Hayes explains why verisimilitude collapses under scrutiny. “Viewing a portrait invokes an unconscious checklist in our visual cortex testing for authenticity. At a glance, AI portraits match enough boxes to trick our brain. But upon extended observation, the uncanny flaws accumulate until the illusion shatters.”

This disparity often stems from the training process. Algorithms are fed thousands of low-resolution web images rather than professional high-megapixel photos. “So while AI portraits look decent at 64 x 64 pixels, when blown up they lack convincing texture and detail,” Hayes states.

Some photographers exploit this flaw for creative expression. Daniel Lee captures public reactions to giant AI portraits displayed in high resolution. “From afar, people are astounded. But when they get closer, you can see them notice and react to the uncanniness. It’s like watching the illusion disintegrate in real time.”

Perfectly Imperfect: Embracing the Quirks and Flaws of AI-Generated Portraits - Analog Nostalgia in A Digital World

In the age of Instagram filters and Facetune, the imperfections of analog photography feel refreshingly authentic. Images captured on film have irregularities like grain, light leaks, and distortions that digital formats lack. This imbues vintage photographs with a sense of nostalgia and humanity that some feel is missing from AI-generated simulacra.

Brooklyn based photographer Maya Levine made waves last year with her series “Analog Soul,” consisting of portraits shot with a 1960s film camera. Despite the retro medium, her images feel surprisingly modern and emotionally resonant. Levine explains her motivations, saying “Digital photography pursuit of technical perfection leaves me cold. The flaws of film capture light and emotion in a way algorithms can’t replicate. Imperfections are what make us human.”

Levine isn’t alone in this perspective. Analogue photography has seen a major resurgence, especially among younger demographics seeking to distinguish their work from increasingly sophisticated AI-generated media. Professional photographer Jordan Moser recently made headlines by shooting his latest magazine spread entirely on 35mm film, a costly and challenging process requiring hands-on craft.

"Digital is too clinical for portraiture," he argues. "The grain and glow of celluloid adds mystery. There are happy accidents in film - double exposures, blown highlights, blurred motion - that create soul AI still can't touch."

This hunger for analog also extends beyond professionals to everyday photography buffs. Vintage Polaroid cameras with their trademark irregular frames have become especially trendy for parties and events. The spontaneity and imperfections of instant film provide a refreshing counterbalance to the "optimized selfies" enabled by AI beautifying apps.

Art and media critic Casey Bright notes, "In the Instagram age, everything is glossy yet sterile in its pursuit of algorithmic perfection. But the flaws of analog processes feel intimately human and can't be faked."

She draws parallels to other fields like music, where listeners crave the "warmth" of vinyl records: "Digital files clinically reproduce sound waves. But the crackles and dynamic range of analog audio involve the listener more viscerally. Similarly, film's visual flaws tug at our heartstrings in a way faultless pixels can't."