Photography and AI

The above image was created with a beta version of an AI generation app that isn’t designed to create photo-realistic images, so we needed to trick it into doing this. The total time from start to finish is under 30 seconds, including the time it took to type the command.

Photography and AI where to now?

This article is for people wondering where photography and AI will fit into our world. The answer is everywhere. Well, that’s my opinion.

A few weeks ago, a well-known photographer commented on social media that we were heading into a new era that would slowly envelop us. I commented below his post that this wouldn’t be slow; it would be like a freight train bearing down on photographers at full speed. How wrong I was! This thing is coming for us at the velocity of an intercontinental ballistic missile—a missile armed with a cluster of nuclear warheads.

At around the same time as the above, there was a lot of talk about a photographer winning a photo competition with an AI-generated photo. Fast-forward a few weeks, and many photo competitions with sections for AI-generated content now exist.

Recently, an article claimed that 25% of people identifying as professional photographers couldn’t pick the difference between AI and camera-created images. A photographer and publisher friend told me it was easy to tell. All you need to do is look at the hands. AI can’t do hands and never will.

Sometimes never, is a very short time. I was recently given an expiring link to a couple of AI images from a research program. For your eyes only, could you have a look and tell me what you think?

Without giving too much away, many people were in the images. Not a dodgy hand amongst them. Detail, lighting, and tonality are all spot on. The expressions were exquisite and believable. The only fault I could spot as a photographer was that there were no faults. They were too perfect.

OK, I get that AI is rapidly developing! But how will it change photography as we know it today?

Let’s look at some of the likely bad outcomes first. Retouching images will become significantly more accessible, allowing unskilled operators to do low to middle-level retouching at home or work. That is a good thing, provided retouching isn’t your job. If it is, you probably have less than 12 months to reskill.

Suppose you are a product or stock photographer creating generic object or location shots for a few dollars on the side. It’s time to find another side hustle. Likewise, designers and graphic artists already feel the pinch from cheap subscription templates and sweat-shop design studios. They will be impacted significantly by affordable and high-speed AI design and graphics solutions.

Another group in the direct firing line are social media influences. Already there are websites offering selfie generation for influencers. You send us a selection of selfies, sign up for a subscription, and we will supply your AI-generated content for your social networks—any location or event you want. We provide your content, photographs and text without you needing to leave home. How long will it take before the internet giants realise cutting out the middleman is more profitable?

Within twelve months, some accounts you love and follow will be AI-generated. End to end, no people involved at all. False names, fantasy personalities and lifestyles. In some cases, that won’t even be a giant step. Already, we have social-media characters faking visits to events and places—pod-casters who allegedly don’t conduct interviews but have themselves edited into the program in post-production.

However, the biggest worry is seeing will no longer be believing. Much of the population currently has trouble picking fake news. What will it be like when believable news images and video flood our lives with lies?

The great advantage of a marriage of photography and AI will undoubtedly be an incredible new toolkit to aid in producing as yet unrealised creative output. I hear cameras able to learn the target, then focus track the subject anywhere in the frame are nearing reality.

Cameras that can wait and watch subjects until they move or behave in a manner of interest are just a little behind—for example, capturing that elusive shot of a bird feeding its young. We currently have two choices: wait, watch and press the button, or set up a movement trigger that captures every movement our bird model makes. With “thinking cameras”, the ability to train the camera only to record when feeding time starts is transformational.