Will A.I. Replace Photographers?
Artificial Intelligence and the magic result of profile pics
We all have seen it. A favourite influencer or close friend posts an A.I. Generated profile pic the likes of which you have never seen.
What amazes you is that they say that all they did was upload a couple of images of themselves at different angles, and artificial intelligence generated numerous new versions of their face.
What is behind all this magic? How is this possible? Such an amazing result for so little effort! Well, it has to do with a concept called machine learning. This machine learning will not only affect how images are created, but it will completely redefine the lives of those creating images. Photographers.
Where did this A.I. photo story start?
This story starts quite a while back more than 60 years ago when the image of the baby below would change photography forever.
Through the use of a rotating drum scanner, Russell Kirsh and his team at the NBS (national bureau of Standards) created a 176 X 176-pixel digital image scan of Russell’s son Walden.
By creating a digital representation of an image, the limits of developing chemicals, photographic paper, and darkrooms would later be a thing of the past. A photo was now a string of zeros and ones representing colours, shades, and hues that algorithms, programs and processors could manipulate. The sky was the limit.
Photoshop – Digital photo manipulation at a new level
Fast forward to 1987 when Thomas and John Knoll developed a photo-processing software called Display, and in 1988, after renaming the software Photoshop, they sold it to a small startup called Adobe Systems.
Through the years, Adobe has continued to update and release new versions of Photoshop with more and more features, such as filters, 3D, and advanced subject selection. Each iteration adds a new level of sophistication and detail to the version of the year before. Adobe Photoshop is a beast of a software!
A.I. Photo software and how photography post-processing is changing
Over the years that followed, Photoshop stayed on top because competitors could not match its high complexity and comprehensiveness. So, with Photoshop’s years of history in photo manipulation, Adobe consistently proved that it was THE feature-rich solution for making digital photos shine.
Therefore, what has kept Photoshop on the top of the heap is that no one was able to match the scope of Photoshop’s capacity to manipulate. No one could do it as well until now.
The increasing number of A.I. Photography solutions. Let machines decide “What is next”
Because Russell Kirsch and his gang were able to make photos a string of ones and zeros, today, we now manipulate photos with comprehensive computing machines.
Manipulate the numbers, and you can manipulate the images.
But the part of A.I. tech that drives this ability is machine learning and, more specifically, deep learning.
Today computers can take an unbelievable amount of data and computing power and develop insight.
What is behind A.I. for photo editing?
We are in a time when photo editing A.I. can analyze all the most captivating photos that ever existed and determine, in every minute detail, all the aspects that make photos more attractive. By doing this, machines can develop the capacity to know a desired income flawlessly and, most importantly, quickly.
New Photo A.I. technologies can even look at every single past action taken to correct, enhance, fix, and adjust a photo, and deduct what needs to be done next to get the desired learned result.
Now, imagine taking this information and applying it to your next photo edit with the press of one button or “prompt”
“The main constraint on innovation is no longer the difficulty in recording and storing information, but the finding of useful insights among the sheer abundance of data now being collected. AI can notice patterns in mammoth data sets that are beyond the ability of human perception to detect. In this way, adopting AI technology can make even mundane and seemingly trivial data valuable.”
Center for international governance innovation CIGI
Most of us are concerned with getting a desired and accurate result. So, RESULTS is the key word! The in-between of a Photoshop edit or technique is a necessary evil that all photo editors have to live through. Learning how to use layers, masking, selection, and other Photoshop techniques has always been the pesky in-between everyone has had to deal with.
The current A.I.-based software has a result-based language that advertises getting rid of the pesky in-betweens.
Midjourney and all its Discord-based iterations can create an image from any text prompt
Luminar and its A.I. editing bundles can do mundane photo edits and make stunningly attractive photos.
Retouch4me does specific professional-level A.I. edits like Dodge and burn, skin healing, eye enhancement and others
Artssy promises a Chat GPT-like experience where you type in a text of a photo you might want and let the software generate its best guess.
Befunky is a savvy social media-based software that does all the collage, skin retouching, and photo optimization you need in one click.
How A.I. makes photos better
A.I. in 2023 does the in-between and the what’s next with more and more style and pizzaz. So, it is a sophisticated form of automation. All of these tools have existed in Photoshop in the form of scripts called actions but with limited results. Whereas now, software like Magicstudio and Luminar is lowering the error margin enough for this automation to be a serious contender.
But how can A.I. change your photography?
The obvious ways A.I. can help photographers
- Doing mundane, time-consuming edits that otherwise take hours to do. Ex) blemish retouching, erasing faults, colour correction etc.
- Generating stylized images from an existing photo to create a new look using the original photo as inspiration
- Guessing where a workflow will go and suggesting a photo edit to finish the job
The less obvious ways A.I. can help photographers
- Photo triage. Making photo choices from a large bank of photos for a wedding client
- Starting entirely from scratch and generating a whole new reality off of a web-based search of images on the world wide web
- Guessing what will be the logical “next photo editing step
Where A.I. doesn’t work too well – Innovation
Here is where A.I. kind of fumbles right now. Past innovations and ideas drive how A.I. makes the next move. Past ideas are an excellent way to base your next decision but not the only way.
- Innovation fuels photography, painting, and music quality. Innovation does not come from always reusing past ideas and decisions.
- All A.I. is limited by where it gets its data. All those at the mercy of these bots will have to live with how the bot makes decisions. This is fine for exact fields but might fall short for more artistic endeavours.
- The brain is not perfect just yet: Errors like Hallucination (the tendency for A.I. to lie or fabricate) and Emergent Properties (Skill learning that software was not supposed to acquire) are still being tweaked.
The Midjourney prompt on this photos was:
A very realistic photo of a hipster photographer holding his camera in a discouraged way he is considering throwing it away – He is in a city environment – ar 16:9 – city is New Your or Los Angeles- image is high definition and high quality 8k
Will A.I. kill Photoshop?
Also, the consensus among photographers is that A.I. Midjourney-generated images have a long way to go before they have any value for the working photographer. Software that helps photographers in their day-to-day is becoming more and more useful, but image generation seems to be a gadget for now.
The Midjourney prompt on this photos was:
A high-definition photorealistic image of a film photography camera exploding on a mound- it is surrounded by other cameras near it and in the background- 8k.
It just did not get that I wanted cameras in the sand around it. It generated little blocks all over the camera, and is satisfied with that. Also, it seems to generate weird letters for the camera brand.
Conclusion – A.I. will change Photoshop
What will change today is the way we use Photoshop and what we expect from the software. Consumers will expect Adobe to take away all the repetitive and mundane actions from Photoshop. The Photoshop interface will be more and more fluid as time goes by.
However, one cannot underestimate that so-called “next steps” of other photography software will become more and more sophisticated as data sets become larger and the computational ability to make sense of the data expands.
I am a photographer, videographer, and copywriter living in Quebec City, Canada. I also have a YouTube channel and an Instagram account dedicated to creation and creativity via my main loves of photography, music, and writing.
To get in touch with me and discuss your collaboration, service needed, or advice, either go to the contact page and write to me via the contact form at the foot of this page.