I’m not against AI. I use it. If I need to clean up glasses glare or swap a background, it does in ten minutes what used to take me an hour of manual masking. As a tool, in the right hands, it’s genuinely good.
What it doesn’t do is replace you. Not your writing voice, not your face on camera, not the thing that makes a client look at your headshot and think “I’d want to work with this person.”
I see the writing version of this constantly. People run a LinkedIn post through AI, paste it in, publish it. The problem isn’t that AI wrote something bad. It’s that AI wrote something generic, and generic isn’t you. The algorithms are getting better at spotting it too. It’s an open secret at this point that LinkedIn quietly promotes AI-flavored content less than it used to. Writing your own words, even clumsy ones, still beats a polished paragraph that sounds like everyone else’s polished paragraph.
There’s a second reason I write my own stuff that has nothing to do with algorithms. I was never a writer. Photography was always the thing I was good at. But the more I’ve forced myself to actually write, the more my own thinking has sharpened. Ideas connect to other ideas mid-sentence that I hadn’t planned on. It’s the same mechanism as brainstorming: the act of generating pulls more out of you than you expected to have.
The same is true behind the camera…
A good headshot session works the same way. I’m not running through a checklist of poses. I’m watching for the half-second where someone’s real expression shows up, the one that happens after they stop performing “professional headshot face” and just react to something I said. That moment is different for every person, because it depends on what I’ve learned about them in the first few minutes: their business, their nerves, their sense of humor.
That’s what an AI headshot generator can’t touch. It’s trained on a huge pile of other people’s photos, and what it hands back is a statistically average version of “professional.” There’s no read on who you are in the room, because there’s no room. Nothing was actually observed. It’s synthetic in the most literal sense.
I’ll say the part most photographers won’t: the AI headshot tools have gotten decent. If someone genuinely has zero budget, using one beats using a five-year-old vacation photo cropped down to a face. But decent isn’t the same as good, and it’s definitely not the same as accurate. It shows people what they might look like on their best-lit, most symmetrical day, generated by a machine that’s never met them. A real session in my SoWa studio, under the constant light panels that I use which creates a soft wraparound light, shows people what they actually look like when the light is doing its job and they’ve stopped tensing up. Those are different products serving different problems.
What a headshot is actually for…
Whoever’s looking at your headshot, a recruiter, a prospective client, someone deciding whether to take your call, is trying to answer one question fast: is this someone I’d trust? That answer comes from something real showing through. A slight asymmetry in a smile. The particular way someone’s eyes engage when they’re genuinely interested in the conversation. AI smooths all of that out because smoothing is what the technology is built to do. Authenticity is built to do the opposite.
That’s the whole case, honestly. Use AI for what it’s actually good at: cleanup, background swaps, the boring technical stuff. Don’t ask it to be you. Not in your writing, not in your photo. The eight faces I’ve included with this post were each shaped by a real conversation between me and the person in the chair, not by a prompt. You can see the difference. Everyone who’s ever hired me could too.
Happy to answer questions if you’re curious how a session actually works.
