How do you make AI photos of yourself? The short answer
To make AI photos of yourself, upload 8 to 15 clear selfies to an identity-trained AI photo generator like MakeAiPhotos, let it train a personal model on your face (15 to 30 minutes), pick a style pack, and generate. You get 30 to 60 realistic photos of yourself in different settings from one upload.
Most guides get this wrong: they send you to Midjourney or DALL-E. Those are text-to-image tools. They cannot generate a photo of you specifically, only a generic person who roughly matches your description. To create AI photos of yourself with your real face, you need a different category of tool.
Two types of AI photo generators, and why the difference matters
Most people searching 'how to make AI photos of yourself' land on the wrong type of tool. There are two fundamentally different categories of AI photo generator, and only one actually creates photos of you.
Text-to-image tools (Midjourney, DALL-E, Adobe Firefly) let you describe a scene and generate a photo, but the person in that photo is not you. They generate a generic attractive face matching your description. There is no way to get your actual face into the output.
Identity-trained generators like MakeAiPhotos work differently. You upload 8 to 15 photos of yourself. The AI trains a personalised model on your specific face: your features, proportions, skin tone, and expressions. Then it generates new photos of you in any setting. The person in every output is recognisably you. (If you are still asking [can AI generate photos of me](/blog/can-ai-generate-photos-of-me) at all, that piece is the direct yes/no walkthrough.)
If you want photos that actually look like you, for LinkedIn, Instagram, or dating apps, you need an identity-trained generator. The rest of this guide covers that process.
What 'AI photos of yourself' actually means
You're not filtering an existing photo. The AI generates a brand-new image of you in a new setting from scratch, it doesn't apply effects to something that already exists. The result looks like a real photograph because technically it is one: a new image rendered from what the model learned about your face.
The output can place you in a professional office, a beach, a European city street, or a studio environment, all from the same set of selfies you upload once.
Step 1, Upload the right selfies
The quality of your output depends almost entirely on what you upload. Good input photos give the AI enough information to learn your face accurately. Poor ones produce results that look like a different person.
What makes a good input photo: clear face visibility with no sunglasses, hats, or hands covering your features; natural lighting from soft window light or outdoor shade; multiple angles including straight on, slightly left, and slightly right; no heavy filters, upload your real face, not an edited version; varied backgrounds mixing indoor and outdoor shots; and recent photos showing your current hair and look.
One tip most guides miss: include at least 2 to 3 photos taken from about 1 metre away, not only arm's-length selfies. Arm's-length shots use wide-angle front cameras that distort facial proportions, making your nose appear larger relative to your ears. Photos from 1 metre away give the AI more accurate proportions to learn from, which shows up in the realism of your output.
Aim for 8 to 15 photos total. Too few means the AI has to guess; too many low-quality ones confuse it. Prioritise variety and quality over volume.
Step 2, Choose your output style
Before generating, decide what you need the photos for. This determines which style to pick.
For LinkedIn and professional profiles: clean studio lighting, business casual or formal attire, neutral or office background. The goal is confident and approachable.
For Instagram and social media: lifestyle settings, coffee shops, outdoor locations, travel vibes. Natural lighting, relaxed expressions, casual but polished.
For dating apps: outdoor lifestyle, casual activities, varied settings that show personality. Avoid anything too staged or obviously studio-lit.
MakeAiPhotos lets you generate across all of these from a single upload. You upload your selfies once and the AI generates photos across multiple styles and settings.
Step 3, Generate, review, and keep the right ones
Once the AI generates your photos, don't just grab the first one. Review the full batch against four checks: Does it actually look like you, eyes, nose, jawline matching your real appearance? Is skin texture natural rather than plastic? Does lighting look like a real photograph with proper shadows? Are hair and background boundaries clean rather than cut-out?
If you get a batch where most results don't look like you, the fix is almost always the input, not the tool. Go back and add more varied, well-lit selfies. Specifically add photos taken from 1 metre away if you only uploaded arm's-length shots.
Save your best 5 to 10. These are your photos.
Step 4, Use them across platforms
One generation session gives you photos for multiple purposes. Update your LinkedIn profile photo, refresh your Instagram grid, add variety to your dating app profile, and use as your profile photo on Slack, email signatures, or team pages. For LinkedIn specifically, see our [AI LinkedIn headshot generator](/ai-linkedin-headshots) and the [professional headshots from selfies](/professional-headshots-from-selfies) walkthrough.
The most effective strategy: mix your AI photos with 1 to 2 genuine photos on each platform. When viewers see AI lifestyle photos alongside a real candid, the AI photos read as professional photography, not AI. The real photos create an authenticity anchor. The AI photos look like you had a professional shoot. This works better than an all-AI profile on every platform.
Selfie-prep checklist: the 10 things to verify before you upload
Before you upload to any AI photo generator, run through this 10-point checklist. Each item below directly fixes one of the top failure modes. Most users skip this step and then blame the tool when results look fake. The fix is almost always in the input.
1. You shot at least 10 selfies, ideally 12 to 15. Below 10, the model does not have enough reference points to anchor your face accurately.
2. At least 2 to 3 of those selfies are step-back shots from one metre away, taken on your rear camera using a self-timer or by another person. The rest can be regular arm's-length front-camera shots.
3. Beauty mode, skin smoothing, face retouching, and any portrait-enhancement filter were OFF in the camera app before you pressed the shutter. Filters cannot be unbaked after the fact.
4. HDR is off and portrait mode is off. Both apply per-pixel processing that softens skin micro-texture in ways that confuse the model.
5. Lighting is soft and directional: window light at a 45 degree angle, open shade outdoors, or overcast daylight. Not harsh overhead office lights, not direct sun on the face, not a dim room with one yellow lamp.
6. You vary the angle across the set: roughly half straight-on, the rest split between slight-left and slight-right turns, plus 1 to 2 three-quarter profile shots.
7. You vary the expression: a neutral baseline, a slight closed-mouth smile, and a full smile. Three expressions across the set is the minimum.
8. No sunglasses, no hats with deep brims, no hair fully covering one eye, no hand covering part of your face. Both eyes visible and in focus on every frame.
9. Background is calm and uncluttered. The face is the dominant subject and fills roughly 60 percent of the frame. Skip group photos and skip anything where you are not clearly the main person.
10. The photos are recent (last 90 days) and match how you actually look right now: current hair length, current facial hair, current glasses. Old photos teach the model an outdated face.
Phone camera settings to change before you start (iPhone, Samsung, Pixel)
Modern phone cameras are aggressively tuned for social media. They smooth skin, lift shadows, and saturate colour on every frame by default. That tuning is the opposite of what an AI photo generator needs to learn your face accurately. Spend 90 seconds turning the following settings off before you shoot a single selfie.
On iPhone (iOS 17 and 18): open Settings, scroll to Camera. Turn Photographic Styles to Standard or set it to a neutral profile with tone 0 and warmth 0. Disable Smart HDR. In the camera app, tap the chevron at the top to expose controls and confirm portrait mode is off, flash is off, and Live Photo is off. Open Settings, Camera, Preserve Settings, turn on Exposure Adjustment so you can pull brightness down by one third of a stop if your face is overexposed.
On Samsung Galaxy (One UI 6 and 7): open the camera, tap the gear icon, scroll to Face Retouching and turn it OFF. Disable Scene Optimizer and Shot Suggestions. Switch off AI features under Advanced Intelligence Features for the camera. Turn HDR to Off rather than Auto. Do not use Beauty Mode or any of the AR Emoji filters.
On Google Pixel: open the camera, tap settings, turn off Top Shot and Frequent Faces. Disable Face Retouching under More Settings. Set Lens Selection to Standard or 1x; do not use the 0.7 ultra-wide front lens. Turn off Night Sight unless you are genuinely in a dark room. Top Shot can replace your shutter frame with a different micro-expression than the one you intended.
Two more settings worth checking on any phone: turn off any third-party camera app filters (Snapchat, BeautyCam, FaceTune passes count as filters even if the photo was saved to camera roll), and check that no Instagram Story filter was applied if you saved a Story photo. If you cannot remember whether a photo was filtered, do not upload it, take a fresh one.
MakeAiPhotos vs other ways to make AI photos of yourself
Most guides ranking for this query confuse two categories of tool. Below is the practical comparison so you pick the right one for your goal. Identity-trained selfie generators are the only category that actually produces photos of you specifically. The rest produce photos of a person who roughly matches your description, which is not the same thing.
MakeAiPhotos (selfie-trained generator): you upload 10 to 15 selfies, the AI trains a personalised model on your face, then generates 40 to 80 new photos of you in packs (LinkedIn, lifestyle, beach, travel, luxury, dating). Total time under 30 minutes for upload and generate, training runs 15 to 30 minutes in the background. Cost is roughly $15. No prompt writing. Output looks like a real photograph of you. This is what most searchers actually want when they search 'how to make AI photos of yourself'. See the [MakeAiPhotos AI photo generator](/generate) and [professional headshots from selfies](/professional-headshots-from-selfies) for the upload flow.
Midjourney, DALL-E 3, Adobe Firefly (text-to-image tools): you type a prompt describing a scene, the AI generates a photo of a generic person matching your description. There is no way to put your specific face in the output, only a face that vaguely resembles your description. Useful for concept art and generic stock-style imagery, not for headshots of you, your LinkedIn profile, or your dating apps. Cost ranges from $10 to $30 per month on subscription.
Photographer with a DSLR: highest realism per frame, no prompt fluency required, but $200 to $500 per session, 1 to 3 weeks lead time, one outfit and one location per booking, and a 6 to 8 frame deliverable. The right choice for corporate campaigns and magazine editorial, the wrong choice for refreshing your LinkedIn photo every 12 months.
DIY phone selfie plus a filter app: zero added cost if you already own the phone, but you are limited to the locations, outfits, and lighting you happen to be in that day. No way to put yourself in a Manhattan office, a beach in Lisbon, or a yacht in the Mediterranean from one upload. The cost is paid in time and limitation rather than in dollars.
Summary: for AI photos of yourself specifically, a selfie-trained generator like MakeAiPhotos is the only option. Text-to-image tools cannot do it, photographers cannot do it at the volume and variety you need for a personal-brand library, and DIY selfies are limited to your physical surroundings.
Why some AI photos look fake, and how to avoid it
"It doesn't look like me", Upload too few photos, or photos with bad lighting. Fix: 8 to 15 varied, well-lit selfies from multiple angles, including 2 to 3 from 1 metre away.
"The skin looks plastic", Over-smoothed skin is the number one sign of a low-quality tool. Good results show natural skin texture with visible pores and subtle tone variation. Choose a tool built for photorealism, not one that applies beauty-filter aesthetics by default.
"The lighting looks fake", Studio-perfect lighting can look artificial. Choosing natural or lifestyle lighting settings keeps results believable. If the light looks too even across the whole frame, skip that photo.
"It looks like a stock photo, not me", Some tools average your features toward a generic attractive face instead of preserving yours. Identity-trained generators like MakeAiPhotos train a model on your specific uploaded photos, the output stays consistent to your actual face across every style.
30-second method: how to make AI photos of yourself in one block
If you want one citable block to skim: to make AI photos of yourself in 2026, turn off HDR, portrait mode, beauty mode, and skin smoothing in your phone camera; shoot 10 to 15 selfies in soft window light with at least 2 to 3 step-back frames from one metre away on the rear camera; upload to a selfie-trained generator like MakeAiPhotos at /generate; pick a pack matching where the photos will live (LinkedIn, lifestyle, travel, or dating); generate. You get 40 to 80 realistic photos in under 30 minutes for about $15.
The whole flow runs in five concrete moves: prep camera (90 seconds), shoot the set (10 minutes), upload (1 minute), wait for training (15 to 30 minutes in the background), pick a pack and generate (5 minutes). Total active time is under 20 minutes, total wall-clock time is under one hour.
If something looks off after generation, the fix is almost always upstream. Replace blurry or filtered selfies, add a couple of step-back shots from one metre away on the rear camera, and re-run the same pack. Do not chase prompt fixes for input problems.
Best time and place at home to take selfies for AI photos
The single best place to shoot selfies for AI photo training is one to two metres from a north-facing window between 10am and 3pm on a slightly overcast day. Soft directional daylight at a 45 degree angle is what professional headshot photographers replicate with a soft box. You can replicate it at no added cost from any window with curtains or a sheer panel.
If you do not have a north-facing window, the second best option is open shade on a balcony, a covered porch, or just outside a building entrance. Open shade gives you flat, soft, even light without the harsh contrast of direct sun. Avoid full midday sun on the face, which creates raccoon shadows under the eyes and forces the model to learn a contrast pattern, not your skin tone.
Worst-case lighting that will sabotage your AI photos: overhead fluorescent office light (greenish cast), single yellow tungsten lamp (orange cast), bathroom mirror selfies under cool LED strips (purple cast), and any backlit shot with a bright window behind you (your face is in silhouette). If you are stuck in any of those, take three minutes to step closer to the largest natural light source in the room and try again.
One more home tip most guides skip: shoot against a plain wall, not a busy bookshelf or bedroom background. A neutral wall (white, beige, light grey, or any single mid-tone) lets the model focus on your face, not on background pattern matching. You can change the background to a Manhattan office or a Lisbon street later in the generated pack.
Common mistakes when making AI photos of yourself (and the fix for each)
Mistake 1: Uploading too many near-duplicate burst shots. Fix: curate down to 10 to 15 frames that show real variety in angle, expression, and lighting. Burst frames teach the model one micro-expression, not your face.
Mistake 2: Only arm's-length selfies. Fix: take 2 to 3 step-back shots from one metre away on the rear camera with a self-timer. The wide-angle front camera distorts proportions; the rear camera at one metre gives true proportions the model needs.
Mistake 3: Leaving beauty filters on. Fix: turn off Face Retouching (Samsung), Photographic Styles plus Smart HDR (iPhone), Top Shot plus Face Retouching (Pixel) before you shoot. Filters bake softened skin into the file and you cannot reverse that later.
Mistake 4: Mixing in years-old photos. Fix: upload only photos from the last 90 days that match how you actually look now (current hair, current facial hair, current glasses). Old photos teach the model a face that is not you anymore.
Mistake 5: Picking a dramatic pack first. Fix: start with a natural lifestyle or business casual pack. Dramatic studio packs amplify every AI artefact in the first run; natural packs hide them and let you check likeness honestly before you commit.
Mistake 6: Reviewing outputs at full screen only. Fix: also review at 96 to 200 pixel thumbnail size, because that is the size strangers actually see your photo on LinkedIn search results, dating-app cards, and Slack avatars. A photo that looks great at 1200 pixels can read as plastic at 120.
Mistake 7: Using Midjourney, DALL-E, or any prompt-only tool and expecting your face. Fix: switch to an identity-trained selfie generator. Text-to-image tools render a person matching your description; only an identity-trained model renders you specifically. For a deeper read see the [professional headshots from selfies](/professional-headshots-from-selfies) walkthrough and the [AI LinkedIn headshot generator](/ai-linkedin-headshots).
Frequently Asked Questions
- How many selfies do I need to upload to make AI photos of myself?
- Upload 8 to 15 photos for the best results. Use varied angles (straight on, slightly left, slightly right), include 2 to 3 photos from 1 metre away rather than only arm's-length, use natural lighting, and skip heavy filters. Quality and variety matter more than total count.
- Will the AI photos actually look like me?
- Yes, if you use an identity-trained generator (not a text-to-image tool) and upload good input photos. Upload 8 to 15 varied, well-lit selfies. If outputs don't look like you, add more photos taken from 1 metre away, arm's-length distortion is the most common cause of likeness problems.
- Can I use AI photos of myself on LinkedIn?
- Yes. AI-generated professional headshots are widely used on LinkedIn. The photo needs to look natural and professional, not obviously AI-generated. MakeAiPhotos generates realistic studio-quality headshots designed for professional use, indistinguishable from photographer-taken results.
- How long does it take to generate AI photos of yourself?
- After uploading your selfies, MakeAiPhotos generates your photos in minutes. You get a full batch of results to review and download, no scheduling, no travel, no waiting days for an edited gallery from a photographer.
- What's the difference between AI photos and just editing a selfie?
- Editing adjusts an existing photo. AI photo generation creates an entirely new image of you in a different setting, outfit, or lighting condition, rendered from scratch using what the model learned about your face. You are not filtering what exists, you are generating a new photograph.
- How do you create AI photos of yourself without a photographer?
- To create AI photos of yourself without a photographer, upload 8 to 15 selfies to an identity-trained generator like MakeAiPhotos, choose a pack (LinkedIn, lifestyle, travel, luxury), and generate. The whole process takes under an hour from first upload to final downloads and costs roughly $15 versus $200 to $500 for a traditional photoshoot.
- What is the easiest way to make AI photos of yourself in 2026?
- The easiest way to make AI photos of yourself in 2026 is to skip prompt-driven tools entirely and use a personal AI photo generator. Upload 10 to 15 clean selfies, pick a pack, click generate. No prompts to write, no Midjourney syntax to learn. The platform handles the photographic structure internally.
- What phone camera settings should I turn off before taking selfies for AI photos?
- Turn off HDR, portrait mode, beauty mode, skin smoothing, and face retouching before you shoot. On iPhone set Photographic Styles to Standard and disable Smart HDR. On Samsung disable Face Retouching and Scene Optimizer. On Pixel turn off Top Shot and Face Retouching. These features overwrite real skin texture, which causes plastic-looking AI output.
- How many sample photos should I upload to make AI photos of yourself?
- Upload 10 to 15 selfies for the best balance of accuracy and variety. Below 10, the model lacks enough reference points to anchor your face. Above 20, near-duplicate frames start hurting consistency. Include 2 to 3 step-back shots from one metre away to give the model undistorted facial proportions.
- What lighting should I use at home for selfies before AI training?
- Stand 1 to 2 metres from a north-facing window at a 45 degree angle, so soft natural light hits one side of your face. If north light is not available, use open shade on a balcony or any indoor window during a cloudy hour. Skip overhead office lights, direct sun, and dim rooms with one yellow lamp, all three teach the model wrong skin tones.
- Can I use Midjourney or DALL-E to make AI photos of myself?
- No. Midjourney, DALL-E 3, and Adobe Firefly are text-to-image tools. They generate a photo of a generic person matching your description, not a photo of your specific face. To make AI photos of yourself, use an identity-trained selfie generator like MakeAiPhotos, which trains a personalised model on 10 to 15 of your uploaded selfies before generating.
- How much does it cost to make AI photos of yourself?
- On MakeAiPhotos, one upload session that generates 40 to 80 AI photos of you costs roughly $15. A traditional photographer session costs $200 to $500 for 6 to 8 retouched frames. The AI option produces more variety from one upload, including outfit, setting, and lighting changes that would require multiple paid sessions to match.
- How do I make AI photos of yourself for LinkedIn versus Instagram versus dating apps?
- Use the same 10 to 15 selfie upload for all three, then generate different packs. For LinkedIn, pick the business casual or professional headshot pack and check our [AI LinkedIn headshot generator](/ai-linkedin-headshots). For Instagram, pick lifestyle, beach, or travel packs. For dating apps, pick outdoor lifestyle and casual activity packs. Mix 1 to 2 real photos in alongside AI photos on every platform.
- How do I take AI photos of myself if I do not have a friend to help shoot the step-back selfies?
- Use your phone's rear camera on a 5 to 10 second self-timer. Set the phone on a stack of books or a low shelf at chest height, frame yourself so the face fills about 60 percent of the frame, tap to focus on your eyes, press the shutter, and step back to where the frame is set. Three step-back self-timer shots cover the gap that arm's-length selfies leave.
- Will AI photos of myself look exactly like a professional photographer shoot?
- For LinkedIn thumbnails, dating app cards, Slack avatars, and Instagram lifestyle posts, the realism gap closes to near zero. For corporate campaigns, magazine editorial, and physical 8x10 prints viewed up close, a professional photographer still wins on per-frame detail. The trade-off is volume and cost: one AI session gives you 40 to 80 frames across multiple settings for $15, versus 6 to 8 frames in one location for $200 to $500.
- How do I make AI photos of yourself that look real, not AI-generated?
- Upload 10 to 15 unfiltered selfies (HDR off, portrait off, beauty mode off), include 2 to 3 step-back shots from one metre away on the rear camera, and pick a natural lifestyle or business casual pack first instead of a dramatic studio one. Review outputs at thumbnail size (96 to 200 pixels) and keep only frames with visible skin texture, matched eye catchlights, and consistent lighting between face and background.
- Can I make AI photos of myself on my phone, or do I need a laptop?
- You can do the entire workflow on a phone. Shoot the 10 to 15 selfies, open MakeAiPhotos in your mobile browser, upload directly from your camera roll, pick a pack, and download to the same camera roll. No laptop, no desktop app, no upload to a cloud drive first. Training runs on our servers in 15 to 30 minutes while you do something else.
- How many photos do I have to upload to make AI photos of yourself?
- The working minimum is 8 selfies; the practical sweet spot is 10 to 15; the ceiling before quality drops from redundant frames is around 20. Below 8 the model lacks enough angle and lighting data. Above 20, near-identical burst photos start biasing the model toward one micro-expression or one bad white balance.
- What is the best AI photo generator to make photos of yourself in 2026?
- The best AI photo generator to make photos of yourself is whichever identity-trained selfie generator nails likeness on your specific face on the first run. MakeAiPhotos is built for exactly this use case: upload 10 to 15 selfies, pick from packs covering LinkedIn, lifestyle, beach, travel, luxury, and dating, and generate 40 to 80 photos in under 30 minutes for about $15. Generic text-to-image tools cannot generate your specific face and are the wrong tool for this job.