I Tried to Create a Model With AI: A Photographer's Honest Take

I've been shooting portraits since the late '70s. My dad gave me a Canon FTB when I was a kid, and I've been chasing light ever since. Over the course of four decades, I've learned how to read a face, work with natural light, and make people comfortable enough to let their guard down in front of a camera.

So when AI image generation started making headlines, I got curious. Could it really replicate what I do? Could it create a believable portrait that looks like something I'd shoot?

I had to find out.

The Experiment

I decided to create a fictional model using AI. Someone who could pass as a real person I might photograph for CreativelyBare or feature on this blog. I wanted to see if the technology could fool even me—someone who's spent a lifetime studying faces, bodies, light, and composition.

I gave the AI specific parameters:

• A woman in her late 20s

• Mixed Asian and Latina heritage

• Natural, girl-next-door energy

• Black and white photography

• Natural window light

• Casual, authentic poses

I named her Maya Chen and even created a backstory: graphic designer, rescue dog owner, loves hiking, makes her grandmother's dumplings on weekends. The kind of details that make a person feel real.

Then I hit generate.

Maya Chen by the window. Except Maya doesn't exist, and I never took this photo. Welcome to AI photography in 2025.

What AI Got Right

I'll admit—the first results were impressive.

The AI nailed certain things:

• Lighting and shadows looked natural and dimensional

• Composition was solid—rule of thirds, leading lines, depth of field

• Skin tones were realistic in black and white

• Facial features were proportional and believable

• The overall vibe felt like something I might actually shoot

If you scrolled past these images quickly on social media or in a blog post, you probably wouldn't think twice. That's how far AI has come.

Where It Falls Apart

But here's where my photographer's eye kicked in.

The more I looked, the more I saw the cracks:

1. The Hands Are A Disaster Hands are notoriously difficult for AI. In multiple generations, I got:

• Fingers that merged together

• Weird angles that don't match human anatomy

• Hand positions that look almost right but feel wrong

• Arms that seem disconnected from the body

Any photographer knows hands are one of the hardest things to pose naturally. AI doesn't understand anatomy—it just averages what it's seen. The result? Hand-shaped blobs that fall apart under scrutiny.

2. Fabric Physics Don't Exist In several images, clothing literally melts into skin. Tank top straps disappear. Material that should drape and fold just... doesn't. The AI treats fabric like it's painted on rather than something with weight and texture sitting on top of the body.

After 40 years of working with natural light on real fabric, these errors jump out at me immediately.

3. The Eyes Don't Quite Connect There's something subtly off about AI-generated eyes. They're technically correct—right size, right placement—but they don't have that spark of presence. That thing that happens when a real person looks at a camera and there's an actual human being behind those eyes.

It's hard to articulate, but you feel it when it's missing.

4. Repetition Without Learning Here's the most frustrating part: every time I hit "generate again," hoping to fix issues, the AI just gave me new problems. It doesn't learn. It doesn't improve. It starts over from scratch each time.

When I work with a real model, we collaborate. We adjust. We problem-solve together. The AI? It's a dice roll every single time.

This is where AI fails: fabric physics. The tank top straps dissolve into skin, the material looks wet instead of cotton, and there's no clear separation between clothing and body. Real photography captures how fabric actually behaves.

What This Means for Photography

So can AI replace real photography?

Not yet. And maybe not ever—at least not for what I do.

Here's why:

Photography is collaboration. It's the energy between photographer and subject. It's the moment someone relaxes and shows you something real. It's adjusting on the fly when the light changes or when you catch an unexpected expression.

Photography is about imperfection. The slight asymmetry in a smile. The way someone's hair falls when they laugh. The genuine moment that happens between poses. AI generates what it thinks is "perfect"—and that's exactly the problem.

Photography is storytelling. Every person I photograph brings their own story, their own energy, their own presence. Maya Chen has a backstory I invented, but no actual history. No nervous laughter before the shoot. No conversation about her life. No real moment captured.

Where AI Actually Helps Me

That said, I'm not anti-AI. I use it constantly.

For my YouTube channel, AI helps me hash out scripts and organize shooting schedules. For blog posts like this one, it helps me structure ideas when my brain won't cooperate. I'd estimate AI contributes to about 50% of my content creation process now.

But generating the actual photographs? The thing I've spent my entire adult life learning to do?

That's still mine.

The Bottom Line

AI image generation is impressive. It's getting better fast. And for certain applications—stock photography, concept art, quick mockups—it's incredibly useful.

But if you want a photograph that captures a real human being? One that has weight, presence, story, and soul?

You still need a photographer.

And you still need a real person in front of the camera.

At least for now.

What do you think? Are you using AI in your creative work? Where do you see the line between tool and replacement? Drop a comment—I'm genuinely curious where others land on this.