The headshot that looks real right up close

AI-generated headshots have crossed a line most people have not clocked yet. At gallery size, on a profile card, they are often flawless. Zoom in, and the cracks appear in the places that only matter when someone actually meets you: the hairline, the ears, the jewellery, the fabric weave. This checklist shows you exactly where to look, and how to fix what you find. It is worth ten minutes before any of these faces goes near your website or your bio.

The test itself takes seconds: open the image, zoom to 300 or 400 percent, and check each of the five zones below. Do every zone every time. Confidence is exactly what these models are good at faking.

Zone 1: The hairline and hair volume

Hair is the tell. Look at where hair meets skin: strands that merge into background, a wisp that flickers, edges that sharpen and dissolve within a few pixels. Then count the hair. Many models produce hair with a slightly dreamy, too-even density, like a wig with perfect lighting. Real hair has flyaways, uneven thickness, and honest chaos at the crown.

Zone 2: The ears

Ears are drawn from memory more than observation, so they drift. Both earlobes should look like the same person’s. Check that earrings, if present, pass through the ear rather than floating near it, and that glasses arms meet an ear that plausibly exists.

Zone 3: The jawline, neck, and collar

The jaw should end where it should and the neck should contain one shadow system, not two. Watch for a jaw that keeps melting into the collar, or a shadow under the chin falling in a different direction to the one on the face.

Zone 4: The eyes and eyebrows

Genuine asymmetry is human; mirror-perfect eyes are not. Check the catchlights, the tiny highlights in each eye, and make sure they come from the same window or lamp. Mismatched catchlights are one of the fastest reveals of a synthetic portrait.

Zone 5: The skin fabric boundary

At high zoom, fabric weave should read as weave, and skin should have pores, or at least an honest effort. The failure to look for is fabric texture fading into smoothness right where it touches skin, or a shirt collar so uniformly soft it reads as illustration.

Fixing what you find

Most faults have a simple remedy. Regenerate the specific zone, rather than the whole portrait, with a sentence that names the fault in plain language: the hairline has gaps where strands merge into the background, add individual flyaway strands, keep the lighting identical. For small structural problems, a targeted edit in a conventional photo editor is often faster than a new render. And the cheapest fix of all is often a crop: if the error lives near an edge, tighten the frame around the face and collar.

The one thing the zoom test cannot fix

Likeness. A headshot of you should survive three seconds of comparison with any recent photo of you. Check the small truths: how your one ear sits, where your part falls, whether that tiny scar or freckle survived the render. Faces carry identity data, models sometimes smooth it away quietly, so trust a friend’s verdict over your own, since you are the least reliable judge of your own face.

Where the five-zone test earns its keep

Zoomed-in scrutiny catches what a swiping audience would never notice, which is the point: the person who matters, the recruiter, a client, the reader of your about page, is exactly the one who pauses right where the render is thinnest. Ten minutes buys a portrait that stays strong wherever someone looks.

Want more like this? Join MOKU Club for free. Weekly resources, early access to new guides, and occasional templates you can actually use. Join below.

When Your To-Do App Becomes the Problem: How PeoplePerHour Freelancers Fix What Automation Can’tDesignHow-toSmall Businesses

When Your To-Do App Becomes the Problem: How PeoplePerHour Freelancers Fix What Automation Can’t

HeraHeraMay 7, 2026
How to Build a Customer Repeat Purchase Predictor That Flags Who Is Due to Order Again Without a PluginDesignHow-to

How to Build a Customer Repeat Purchase Predictor That Flags Who Is Due to Order Again Without a Plugin

HeraHeraOctober 15, 2026
AI for Social Media Comment Moderation: How Small Businesses Filter Spam and Troll Comments Across Every Platform Without Checking Each One ManuallyAiTips & Tricks

AI for Social Media Comment Moderation: How Small Businesses Filter Spam and Troll Comments Across Every Platform Without Checking Each One Manually

HeraHeraOctober 7, 2026