Seeing People, Not Pixels: AI and Best Practices from a Photographer’s Perspective

AI-generated image. This is not real.

To create this effect, I uploaded a photo of myself and “prompted” ChatGPT to create a photorealistic image of a community-based photojournalist using a Canon camera.

AI is everywhere in generated portraits, avatars, and synthetic scenes appearing on websites, social media feeds, and even in televised advertising.  

As a Marietta photographer specializing in nonprofit storytelling photography, product and portrait photography, I’ve been thinking carefully about where I stand. 

This isn’t a rejection of AI, but rather a conscious choice about how I use it. AI can be creative, fast, and even resource-efficient. It’s great for branding illustrations, conceptual visuals, or playful identity projects.

But photography tells a different story: real people, real moments, and real experiences require presence, trust, accountability, and the ability to capture the ‘decisive moment.’

When I photograph nonprofit events, community initiatives, or portraits, I’m documenting lived experiences that carry ethical, emotional, and historical responsibility. AI can’t replace that, and it shouldn’t try.

Other HUGE considerations matter, too: energy and water usage at scale, and privacy concerns for clients and communities. Thoughtful use is essential.

Here’s my approach:

• Use AI for creative exploration and branding projects

• Use photography for real stories, nonprofit work, and portrait photography

• Generate fewer, purposeful images

• Respect privacy, consent, and authorship

Photography remains an act of witnessing, a way to honor presence and preserve truth. AI and photography can coexist—but intention matters.

As a Marietta photographer, my focus is to be mindful of authenticity, people, and care about the stories I capture and the impact they leave behind. See the energy and water use estimates below for image making using AI:

Energy Use for AI image generation:

• Generating one AI image through ChatGPT/DALL·E: ~0.05 kWh on average.

  That’s roughly the energy to charge a smartphone 4–5 times.

• Generating 1,000 images: ~50 kWh, similar to running a standard fridge for 2–3 weeks.

Water Use for AI image generation:

• Most data centers use water to cool servers, averaging about 1.5 liters per kWh.

• For one image: 0.05 kWh×1.5 L/kWh=0.075 L, about 75 milliliters, or half a cup of water.

• For 1,000 images: 50 kWh×1.5 L/kWh=75 liters or roughly enough water for 2–3 ten-minute showers.