AI in Photography – Good or Bad?

Over the last few months, the use of AI in photography has been subject to much discussion and the inevitable argument on social media. It is one of those issues that divide opinion. Do we really know what we mean when we talk about AI in photography? Is it really going to be the death of picture-taking? While I don’t profess to be an expert, I’m open to trying to understand more to see if we need to be as scared as some say.

What is AI?

According to Encyclopaedia Britannica, the definition of AI is the ability of a computer, or a robot controlled by a computer, to do tasks that humans usually do. A phrase which has been associated with AI is machine learning. This is when a computer is trained to learn from its inputs but without explicit programming for every circumstance.

In the case of photography, AI has led to several advances that have changed how images are produced. They can be broadly placed into two buckets. The first is assistive, where the computer will help to make a task easier. Secondly, generative is where the computer will create a photograph based on a phrase or description. Let’s look at how these two types of AI can impact photography.

Assistive AI

If you use the camera on your mobile phone, you probably use AI without knowing it. The blurry background in ‘portrait mode’ isn’t created by the usual physical elements such as aperture or distance. Instead, it uses computational photography to determine the subject and then applies blur to everything else.

The latest mirrorless cameras from Canon, Nikon and Sony have all introduced amazing auto-focus systems. Settings have been introduced to automatically focus on people, animals, and cars. The processor in the camera has been taught by analysing thousands of images to be able to lock on to the subject.

When we come to edit our photos, software developers have used machine learning to improve the functionality of tools. In Adobe Photoshop and Lightroom, you can select subjects in seconds and then apply specific changes to that part of the photograph.

It’s difficult to argue that these AI features aren’t helping us to take better photographs. We still have a high degree of creative control, but technology is helping us to achieve better results more often. However, the next stage is where we may start to raise some of the heated arguments seen on social media.

Suggestive AI

When I first started writing this section, I used the heading Overly Assistive AI. It was probably a little judgmental, so I will call this next use of AI suggestive. This is where the machine-learned tech will take control of your camera and suggest (or automatically apply) settings to get the ‘perfect’ shot.

One of the most funded Kickstarter photography gadgets is a tool called Arsenal. It’s described as an intelligent camera assistant. It claims to be able to examine a scene while you are at the location and suggest the settings for your camera to get a perfect shot. It measures the stability of your camera and suggests an appropriate shutter speed. It can decide when to use multiple exposures to create high dynamic range shots. It even claims to be able to remove people from scenes by taking multiple images with enough information to ensure it can clone them away. You decide – is this taking the personal creativity out of photography by essentially giving you an intelligent auto-exposure functionality? You can ignore the suggestion, but will you if you have spent $320 on a tool that takes and edits your pictures?

Luminar AI is a popular editing tool. As it suggests in the name, much of the functionality in Luminar is based on machine learning. One of the features offered is an automatic composition tool. With one press of a button, Luminar will decide what the best crop is for the scene taken. You can decide for yourself below how well it does. The image on the left is the original composition, unedited. The one on the right is the automatic composition applied in Luminar AI.

The discussion with these types of tools is what do we mean when we say we want the perfect shot? We can look at some of the most iconic images taken, and they don’t always have the conventionally ‘best’ settings.

Generative AI

This last use of AI has proved to be the most controversial. This is when an image can be generated from writing a phrase or description. If you want to see an example in action, look at the top of this blog. I don’t have a robot that can take photographs in my garage. Even if I did, I don’t have the time or weather to photograph it in front of a Scottish mountain. However, by typing ‘a computer taking a photograph of a Scottish mountain’, I now have a photograph doing precisely that!

There are definitely issues for the photography industry with this technology. The standard of output is OK, but I wouldn’t put it equal to a professional photographer. That will change, however. I can easily see quality results being available in the next few years.

I can see a challenge to photographers in the future with the use of AI, but what about the current photographers and even those of the past? The images generated from the line of text are gathered from other photographs. Some elements of a generated photograph will inevitably be taken from part of an existing photograph. This leads to the question of copyright. If a substantial part of a generated image comes from another photograph, what attribution is given to the original photographer or artist? I have reverse-searched the header image but have been unable to find either the robot or the mountain. However, in December 2022, an AI image generator, Lensa, was called out for inserting fragments of artists’ signatures in its generated images. The company argued that the machine had learnt that signatures formed part of drawn images and created a signature without referring back to its original database of images. Who knows if this is true?

What do you think?

Is AI going to help or hinder the photography process? Are there elements of AI you are comfortable with and others where you think there should be controls? I’d be interested to see your views in the comments section.

Please give us your feedback.

If you’ve got any questions or comments, leave them below. You can sign up for the Edinburgh Photography Workshop monthly newsletter, where you’ll get regular updates on exciting things happening in photography and some great tips. Sign up by clicking here.

About the author

As well as running Edinburgh Photography Workshop, Rich Dyson is a professional photographer. His photographs are regularly used in newspapers such as The Times, Guardian and Daily Telegraph. He also had two solo exhibitions and was featured in a members-sponsored exhibition in the Scottish Parliament. You can see and buy his photography at richdysonphotography.com.