Tutorials12 min read
Why Your AI Instagram Photos Look Fake (And How to Fix It)
AI photos get read as fake in a feed for four specific reasons — phone focal lengths, unnamed light, over-polish, dead composition. Each has a one-line fix.
On this page
- The problem is not your prompt list
- What gives an AI photo away in a feed
- Fix 1: Name the camera, not the vibe
- Fix 2: Give the light a direction and a flaw
- Fix 3: Write the pose as an action, not an adjective
- Fix 4: Fill every slot instead of writing a nicer sentence
- Copy-ready skeletons for the three feed formats
- More Instagram-style prompts to start from
- FAQ
- Why do my AI photos look fake on Instagram?
- What aspect ratio should I generate for Instagram?
- Which model is best for realistic Instagram-style photos?
- Should I write selfie prompts as JSON?
- Can I use my own face?
The problem is not your prompt list
People can tell. An AI photo can be technically flawless and still get read as fake in half a second of scrolling, and adding more style words — "instagram aesthetic, 4k, highly detailed" — makes it worse, not better.
The reason is specific and fixable: feed photos carry evidence of the device and the moment that made them, and most prompts erase that evidence. This article is about the four lines that put it back, checked against what 9,599 published prompts in the LocalBanana gallery actually contain.
What gives an AI photo away in a feed
Three tells, in the order people notice them.
1. Light with no source. The image is evenly lit from everywhere. Real phone photos have one dominant light and everything else falls off — a window on the left, a ceiling bulb overhead, a flash that blows out the nearest shoulder. In our corpus, the single most-requested lighting term is studio lighting at 6.5% of all prompts, ahead of natural light (3.8%) and window light (1.1%). The most-asked-for light is the least feed-native one.
2. No camera identity. This is the biggest one, and the corpus shows it clearly:
4.9%
of prompts specify 85mm
The portrait-studio focal length
0.75%
of prompts specify 24mm
The focal length phones actually shoot
45.5%
of Nano Banana prompts carry photographic parameters
vs 28.0% for GPT Image, from the same 9,599-prompt corpus
A phone's main camera sits around 24mm equivalent. That wide field of view, with its slight edge stretch and deep-ish focus, is the visual signature of every casual photo you have ever scrolled past. Almost nobody asks for it. They ask for 85mm — which produces a compressed, creamy, studio look that no one's phone has ever produced by accident.
3. Zero imperfection. Nothing is cut off, nothing is blurred, the hands are perfect, the subject is centred. Real feeds are full of half-closed eyes, a knee cropped by the frame edge, motion smear from someone walking past.
What 'aesthetic' does to a prompt
Words like "aesthetic", "beautiful", "stunning" and "4k" push the model toward its average of polished stock imagery — precisely the look that reads as synthetic. Every fix below replaces one of those words with a fact.
Fix 1: Name the camera, not the vibe
Write the device and the focal length. This one change does more than any style keyword, because it constrains geometry rather than mood.
Stairwell Glance
Nano Banana ProPrompt▾
{
"meta": {
"quality": "ultra photorealistic",
"resolution": "8k",
"camera": "iPhone 15 Pro Max",
"lens": "24mm",
"aspect_ratio": "9:16",
"style": "dim stairwell lighting, raw iphone candid"
},
"scene": {
"location": "apartment stairwell",
"time": "night"
},
"subject": {
"action": "walking upstairs",
"pose": {
"one foot higher step": true,
"looking back": true,
"expression": "playful"
},
"outfit": {
"top": { "type": "tight crop sweater", "color": "grey" },
"bottom": { "type": "micro skirt", "color": "black" },
"legs": { "type": "black thigh-high stockings" }
}
},
"camera_perspective": {
"pov": "few steps below",
"angle": "slightly low"
}
}Notice what is in that prompt and what is not. It specifies a lens, a location, a time of day and a lighting condition. It does not contain the word "aesthetic". The look falls out of the constraints.
The models are not equal here. Across the corpus, 45.5% of Nano Banana prompts contain photographic parameters, versus 28.0% for GPT Image and 9.6% for Midjourney — people have already sorted themselves, sending camera-language work to Nano Banana. If you want feed-realistic people photos, start there and browse the Nano Banana prompt collection for working examples.
Three camera lines that cover most feed situations:
Shot on iPhone 15 Pro, 24mm main camera, handheld, slight lens distortion at the frame edges
Shot on a compact digital camera with direct on-camera flash, harsh falloff, mild sensor noise
Shot on a phone front camera at arm's length, 9:16, subject slightly off-centre, edges soft
Fix 2: Give the light a direction and a flaw
"Soft lighting" tells the model nothing about where the light is. Name the source, the side it comes from, and what it fails to reach.
Morning selfie with platinum hair
Nano Banana ProPrompt▾
{
"prompt_type": "Ultra-Photorealistic Image Generation",
"subject": {
"demographics": "Young woman, roughly 20s",
"hair": {
"color": "Platinum blonde",
"style": "Long, straight to slightly wavy, loose and fanned out on the bed",
"texture": "Silky, fine strands, realistic stray hairs visible"
},
"skin": {
"tone": "Fair/Pale with warm undertones",
"texture": "High fidelity skin texture, visible pores, natural skin sheen, slight flush on cheeks",
"details": "Hyper-realistic skin imperfections"
},
"eyes": {
"color": "Light (blue or grey)",
"makeup": "Long defined eyelashes (mascara), subtle eyeliner",
"gaze": "Looking slightly off-camera to the side, soft and dreamy"
},
"mouth": {
"lips": "Full, natural pink shade, slight gloss or balm finish",
"expression": "Relaxed, neutral to slightly parted lips, moody pout"
}
},
"attire": {
"clothing": "White ribbed tank top/camisole",
"fit": "Tight-fitting, scoop neckline showing décolletage",
"fabric_texture": "Soft cotton ribbing details visible"
},
"accessories": {
"neck": "Thin silver chain necklace with a small cross pendant resting on the chest",
"ears": "Small silver hoop earring visible on the left ear"
},
"pose_and_composition": {
"position": "Lying down on back (supine)",
"angle": "High-angle selfie perspective (point-of-view), close-up on face and upper torso",
"framing": "Diagonal composition, head tilted to the right"
},
"environment": {
"setting": "Bedroom, lying on a bed",
"background": "White bed sheets/duvet",
"texture": "Crumpled white fabric, soft folds, linen texture"
},
"lighting": {
"type": "Soft diffused indoor light or window light",
"quality": "Soft shadows, gentle highlights on the forehead, nose bridge, and chest",
"mood": "Intimate, morning vibe, soft glow"
},
"styling_and_mood": {
"aesthetic": "Y2K selfie, girl next door, candid, intimate, 'coquette' subtle undertones",
"atmosphere": "Relaxed, alluring, lazy morning, authentic social media snapshot"
},
"camera_technical": {
"style": "Smartphone photography or compact digital camera",
"lens": "24mm wide-angle (simulating phone camera)",
"aperture": "f/2.2",
"focus": "Sharp focus on eyes and lips, slight depth of field blurring the hair edges",
"resolution": "8k, RAW photo, ultra-detailed",
"processing": "Natural color grading, slight desaturation, realistic contrast"
}
}
Boyfriend POV — A Quiet Gaze
Nano Banana ProPrompt▾
A photorealistic vertical photo (9:16) captured from a boyfriend point of view. Medium close-up composition, the subject occupies roughly 60% of the left side of the frame, with a slightly top-down angle, as if taken casually by someone sitting across the table. A light-colored round wooden table edge appears in the lower-left foreground. The subject is a young Asian woman with short brown hair and soft bangs. She wears a black half-zip high-neck fleece sweatshirt, thick and soft in texture. A black square patch with white “alo” lettering is clearly visible on her left upper sleeve. She leans forward naturally, left elbow resting on the table, hand supporting her cheek or chin. She looks directly at the camera with a gentle, relaxed expression. The background is a busy but blurred coffee shop interior: industrial exposed ceiling pipes, track lighting, a distant service counter with a red menu board, and softly blurred customers in dark clothing. Shallow depth of field keeps the subject sharp and the background creamy and soft. Camera feel: 50mm or 85mm prime lens Wide aperture (f/1.8 – f/2.0) Warm indoor lighting Natural skin tones Overall mood: candid, intimate, everyday, a spontaneous lifestyle moment typical of social media boyfriend POV photography.
The scarcity table is the opportunity. Out of 9,599 prompts, window light appears in 1.1%, overcast in 1.5%, side lighting in 0.5%. Those are three of the most feed-native lighting conditions in existence and almost nobody asks for them, while 6.5% ask for a studio.
Single window light from camera left, warm bounce on the shadow side, no fill
Overcast afternoon, flat soft light, no visible shadows, slightly cool white balance
Direct on-camera flash at night, subject bright, background falls to near black within a metre
Fix 3: Write the pose as an action, not an adjective
"Confident pose" produces a catalogue stance. A verb produces a photograph.
Tokyo Night Candid Capture
Nano Banana ProPrompt▾
Ultra-realistic candid smartphone photograph, 9:16 vertical. Nighttime on a quiet Tokyo city sidewalk beside a building with white ceramic tile walls. The subject is walking past the camera, partially side-on, captured mid-step in a rushed, accidental moment. The subject reacts instinctively to being photographed: turning her head back mid-walk as if surprised, raising one hand close to the lens to partially block the camera, shy, restrained smile forming as she suppresses laughter, body angled in half-profile while continuing to walk forward, motion carries her slightly out of frame, imperfect framing. Shot on a modern smartphone, handheld, rushed capture. Strong uncontrolled camera shake. Heavy directional motion blur across the raised hand, hair, and face. Flash-only lighting fired mid-motion. Harsh flash highlights on skin and hand. Deep surrounding darkness. Extremely noisy. Heavily blurred. Raw smartphone look.
Kawaii Cafe Finger Heart Selfie
Nano Banana ProPrompt▾
{
"portrait_prompt": {
"subject": {
"description": "Based on <User Portrait>, a young woman with playful cute expression, both hands making mini finger heart gestures near her face, pouty duck lips, head slightly tilted, looking up at camera with aegyo charm",
"features": {
"hair": "hair with subtle highlights, light blue snap hair clip on left side",
"eyes": "Big sparkling doe eyes looking upward at camera, soft brown eyeshadow",
"expression": "Playful pouty duck face, slightly pursed coral lips, rosy pink blush on cheeks, mischievous cute vibe",
"makeup": "Korean ulzzang style - emphasized rosy blush, gradient coral lips, natural brows, dewy skin"
},
"pose": "Seated at cafe table, both hands raised near face level making mini finger heart gestures (thumb and index finger tips touching), elbows resting on table, shoulders slightly raised in cute shrug"
},
"attire": {
"type": "Based on user-uploaded clothing reference",
"details": [
"Light blue cable knit oversized sweater",
"Wide boat neckline exposing collarbones",
"Chunky knit texture visible",
"Silver chain necklace with small pendant",
"Multiple delicate rings"
],
"color": "Soft baby blue"
},
"composition": {
"shot_type": "Medium close-up, upper body portrait",
"focal_length": "35-50mm",
"camera_angle": "Elevated overhead angle, camera positioned above eye level looking down at subject, boyfriend/girlfriend POV selfie angle",
"camera_height": "Above subject's head height, shooting downward",
"framing": "Subject centered, table surface and props visible in lower portion of frame"
},
"lighting": {
"type": "Soft diffused natural window light",
"direction": "From right side and slightly behind, wrapping around face",
"mood": "Bright airy cafe lighting, soft shadows under chin, even illumination on face"
},
"color_palette": {
"film_simulation": "ASTIA",
"style": "Fujifilm ASTIA inspired color grading, soft creamy skin tones, pastel color harmony, low contrast, gentle highlight rolloff, slightly desaturated dreamy feel",
"tones": ["Baby blue", "Cream white", "Soft pink", "Pastel mint accents"]
},
"environment": {
"setting": "Kawaii aesthetic cafe with anime/Popmart collectibles",
"background": "White wooden panel walls, pink neon sign (blurred 'something' text), doorway/mirror reflection showing cafe interior",
"table_props": {
"left_side": "Sailor Moon anime figurines - Sailor Venus (blonde) and Sailor Mercury (blue hair) in sailor uniforms, standing poses",
"center": "Matcha latte with heart latte art in clear glass cup on turquoise saucer",
"right_side": "Small blue Popmart-style toy figure (BT21 Koya koala character)"
},
"atmosphere": "Cute otaku cafe date aesthetic, collectible figurines display"
},
"mood": "Playful aegyo, sweet mischievous charm, Korean cafe date selfie vibe, girlfriend material",
"technical_tags": [
"8K",
"photorealistic",
"overhead angle",
"shallow depth of field",
"soft skin texture",
"natural beauty",
"instagram aesthetic"
],
"character_reference": "Based on user-uploaded character reference",
"clothing_reference": "Based on user-uploaded clothing reference",
"negative_prompt": "Eye-level shot, looking down, harsh shadows, orange skin, heavy contour, serious expression, flat angle"
}
}The pattern: describe what the body is doing at the instant of the shutter. Walking past and turning back. Reaching for a cup. Pushing hair out of the eyes. Laughing at something outside the frame. Each of those forces asymmetry and a little blur — the two things a posed AI portrait never has.
✕ Bad
beautiful girl in a cafe, instagram aesthetic, golden hour, 4k, highly detailed, professional photography✓ Good
Candid photo, iPhone 15 Pro 24mm, young woman mid-laugh reaching for a coffee cup, head turned away from the lens, single window light from camera left, plain white tiled wall behind, slight motion blur on the hand, 4:5Fix 4: Fill every slot instead of writing a nicer sentence
This is the strongest pattern in the whole corpus. 13% of published prompts (1,249 of 9,599) are written as structured JSON rather than prose. Their median gallery views are 98, against 9 for prose — and the gap survives controlling for length: in the 80–149 word band it is 141 versus 17.
What that number is and is not
Views are influenced by how long an entry has been published and where it landed in the feed, so this is a correlation, not a controlled test. The mechanism worth borrowing is mundane: a labelled object makes an empty slot visible. You cannot leave lighting blank without noticing, whereas a nice prose sentence can run start to finish without ever saying where the light comes from.
High-angle Chic Influencer Portrait
Nano Banana ProPrompt▾
{
"subject": {
"description": "A stunning high-angle shot of a chic Asian fashion influencer with a cool, alluring attitude.",
"age": "20s",
"expression": {
"eyes": {
"look": "sharp fox-eyes, piercing gaze directed at camera",
"energy": "confident, slightly cold, seductive",
"details": "defined eyeliner, emphasized aegyosal"
},
"mouth": {
"position": "relaxed lips, subtle smirk",
"energy": "chic"
},
"overall": "stunning, high-visual-impact beauty"
},
"face": {
"preserve_original": false,
"makeup": "high-contrast makeup, pale porcelain skin, reddish gradient lips, sharp jawline, V-shape face",
"style": "cool-toned beauty, K-pop idol visual"
},
"hair": {
"color": "black",
"style": "long sleek straight hair with full straight bangs",
"effect": "glossy, high-fashion finish"
},
"body": {
"frame": "slim, petite, fragile aesthetic",
"pose": {
"position": "leaning forward significantly",
"overall": "dynamic foreshortening, emphasis on head and upper torso"
},
"skin": {
"tone": "cold fair skin",
"lighting_effect": "brightened face, soft beauty lighting, no dark shadows"
}
},
"clothing": {
"top": {
"type": "ultra-fine gauge knit top",
"color": "cool grey",
"details": "mock neck, skin-tight fit, lightweight thin fabric (not thick)",
"effect": "perfectly sculpting body curves, smooth texture"
},
"bottom": {
"type": "dark pencil skirt",
"details": "high waisted with thin luxury belt"
}
}
},
"photography": {
"camera_style": "High-end social media snapshot",
"angle": "High angle POV",
"shot_type": "Medium close-up",
"aspect_ratio": "9:16",
"texture": "clear, sharp, slightly filtered for beauty",
"lighting": "overcast cool daylight, soft diffuse light"
},
"background": {
"setting": "European classic architecture",
"atmosphere": "fashionable street corner",
"blur": "bokeh background to emphasize subject"
},
"negative_prompt": [
"round face",
"plain face",
"no makeup",
"warm yellow skin",
"chunky knit",
"thick sweater",
"loose clothing",
"wrinkled fabric",
"dull eyes",
"friendly boring smile",
"low resolution",
"dark lighting"
]
}You do not need strict JSON syntax. A labelled list does the same job:
subject: young woman, mid-20s, shoulder-length dark hair, no makeup
wardrobe: oversized grey hoodie, silver hoop earrings
setting: convenience store aisle, late night, fluorescent ceiling tubes
camera: iPhone 15 Pro, 24mm, held at chest height, slight upward tilt
light: overhead fluorescent only, greenish cast, hard shadows under the eyes
action: turning toward the lens mid-step, one hand on a shelf
imperfection: slight motion blur on the hand, frame tilted a few degrees
format: 4:5 vertical
Copy-ready skeletons for the three feed formats
Swap the bracketed slots. Everything else is doing structural work — leave it in.
Mirror selfie
Ultra-realistic mirror selfie taken on [PHONE MODEL], 24mm wide lens, 4:5 vertical.
Subject: [AGE + APPEARANCE], wearing [OUTFIT].
Location: [BATHROOM / BEDROOM / GYM], [TIME OF DAY].
The phone is visible in the reflection, held at chest height, screen glow on the hand.
Light: [SINGLE SOURCE + DIRECTION], no fill, visible falloff on the far side of the face.
Imperfection: fingerprints and light dust on the mirror, frame tilted slightly, mild sensor noise.
No beauty filter, natural skin texture with visible pores.
Candid street or café shot
Candid photo, shot on [PHONE MODEL], 24mm main camera, 4:5 vertical.
Subject: [PERSON], caught mid-[VERB], [WHERE THEY ARE LOOKING].
Setting: [PLACE], [TIME OF DAY], [WEATHER].
Light: [SOURCE] from camera [LEFT / RIGHT / BEHIND], background falls off darker.
Framing: subject off-centre, part of [OBJECT] cropped by the frame edge.
Motion: slight blur on [MOVING PART], everything else sharp.
Colour: unedited phone JPEG look, no film emulation.
Night flash portrait
Direct on-camera flash photo at night, compact digital camera look, 4:5 vertical.
Subject: [PERSON + OUTFIT], [ACTION], reacting to the flash.
Setting: [PLACE], after dark.
Light: hard frontal flash, bright on the nearest surfaces, background drops to near black within a metre.
Artefacts: visible sensor grain in the shadows, slightly blown highlights on the forehead.
Composition: shot from slightly above eye level, subject not centred.
More Instagram-style prompts to start from
Each of these is a published gallery entry with its full prompt attached — open one, copy it, change the subject, and run it in the generator.
FAQ
Why do my AI photos look fake on Instagram?
Almost always three things at once: the light has no direction, the prompt names no camera, and the pose is described with an adjective instead of a verb. Fix the camera line first — writing "iPhone 15 Pro, 24mm" instead of leaving focal length unspecified changes the geometry of the whole frame. Only 0.75% of the prompts in our corpus specify 24mm, which is part of why so much AI output looks like it was shot on an 85mm portrait lens in a studio.
What aspect ratio should I generate for Instagram?
4:5 for the feed — it is the tallest portrait crop the feed supports, so it occupies the most screen. 1:1 if you want a predictable grid. 9:16 for Stories and Reels. Generate at the ratio you will post at rather than cropping later; the composition changes when you crop.
Which model is best for realistic Instagram-style photos?
Nano Banana, on the evidence of how people actually use these models: 45.5% of its prompts in our corpus carry photographic parameters, against 28.0% for GPT Image and 9.6% for Midjourney. That is revealed preference across 9,599 published prompts, not a benchmark — but the split is large and consistent. The full breakdown is in what 9,667 published AI prompts actually contain, and the head-to-head is in Nano Banana vs DALL·E.
Should I write selfie prompts as JSON?
The syntax is not the point; the labelled slots are. Structured prompts in our corpus show far higher median engagement than prose at the same length (141 vs 17 in the 80–149 word band), and the plausible reason is that a labelled list makes a missing specification visible. A plain list with subject: / light: / camera: labels captures most of the benefit.
Can I use my own face?
Yes — many gallery prompts are written for it, using a placeholder such as Based on <User Portrait> and then specifying that facial features must be preserved exactly. Upload your reference in the generator and keep the identity-preservation line in the prompt; that instruction is what stops the model from drifting into a generic face.
LocalBanana Team
We run Nano Banana, GPT Image and other models side by side, and publish every prompt in our gallery. These guides are written from that corpus.
Updated @LocalBanana_io
