AI Has Officially Destroyed Commercial Photography

AI image generation has been incredible for the last year, but it was still extremely difficult to generate realistic images with specific subjects, people, or products. Google’s new AI model can finally do it, marking the end of the commercial photography industry. 

As artists, we like to pretend that companies respect our work; that they are willing to pay a premium for our photos because of some intrinsic artistic value, but in reality, it’s simply a business expense. Photography, or imagery in general, is a small line item on the marketing budget. Companies do not care who took the picture, if it was shot on film, if it required complex lighting, if it was Photoshopped, or if it was generated by AI, because their customers don’t care. The pictures simply need to catch a potential customer’s attention and convince them to buy. 

The stock photography market was already decimated years ago by microstock websites that charged less than $1 for royalty-free images, but why pay a dollar for a photo that has been used hundreds of times when you can generate a better-looking custom image with AI?

Midjourney, my favorite AI image generator, costs $30/month, but why spend that when every large LLM like Chat GPT, Grok, or Gemini will also generate images completely for free? Stock photography was already almost worthless, but now it’s completely worthless. Anyone on earth can now generate what would have been a world-class photo, instantly, and for free. 

Until now, commercial photographers could still compete with AI because companies needed imagery of specific people, places, or things. AI models could create a generic image of “headphones,” but they couldn’t create “our headphones.” AI could generate a photo of a “girl wearing a dress,” but they couldn’t make her wear “this dress.” That is until now. 

Last week, Google released its new AI image generation model, Nano Banana AKA Gemini 2.5 Flash. As far as I can tell, this model is completely free. If you use it for hours like I did, you’ll run out of “tokens,” and they will lock you out for a day, but if you switch your Google account, you can keep generating. 

This new model can finally do what we have all been afraid of; it can keep subjects consistent image to image. You can generate a person, place, or thing from scratch and then make small changes to the image without changing the subject, or you can upload your own images, and Gemini will create new photorealistic images using a real subject. This sounds too good to be true, but I’ve been playing with it for days, and it’s incredible. Gemini generated all of these images of my hot sauce brand, Oliveum, in about 15 seconds after I uploaded a real image of the bottle and box. 

Literally the only downside is that these files appear to be locked at 2MP, which is fine for social media but not quite enough for print. Midjourney can currently create 4MP images, and I’ve run those files through AI upscaling software, and I can get those files to look good enough for print. In a few months, Google will certainly unlock a higher resolution option. 

With this model’s release, AI has made almost all commercial photography obsolete, and it will only get better from here. It’s depressing to have spent my entire adult life working on a craft that appears to have become worthless overnight, but there is no point in complaining or attempting to fight against this; it’s already happened. 

I’m going to do a follow-up video about this and my predictions for the future, but right now I’d like to hear your thoughts in the comments below.