Has Ai Killed Art? –

Ai or Artificial Inteligence/Machine Learning has been around for a while now, but never has it’s presence been more keenly felt than in and around the Creative Arts. Producing imagery was once the domain of illustrators and graphic designs.  Now it appears that everyone with a computer/tablet/smartphone and a connection to Dalle and Midjourney can produce the most amazing images.

At first as a photographer I didn’t feel particularly threatened by ai.  I suspected that what I was producing was safe for a while and that everyone would always need good photographers?  As a simple test I ran a text prompt through Midjourney  – a text to image ai generator. The results to me seemed less than impressive.  Below is an example of what I do and what ai thinks I do.  To generate the image below I used the text prompt “Oil on Water”.

I decided to give the ai a second attempt.  Although the result was better, it was still significantly different from the kind of imagery that I was producing.  Below is the second time I ran the prompt ‘oil on water’.

Fortunately or unfortunately depending upon a person’s perspective, technology forges ahead. Shortly after producing the ai images above, the original makers of Dalle  launched Dalle 2. This software built upon the original concept of the text to image prompt, but with two significant developments. Firstly there is now the ability to add your own images and see how the ai reinterprets them. This is done by uploading an existing image, running the software and seeing what it comes up with. Secondly, there is also the ability to upload an image and remove parts of it to allow the ai to fill in the blanks.  At first, removal of faces was disallowed because this was too close to the notion of ‘deep fakes’.  However this temporary block has now been lifted and there are now many examples of pop star faces being added to different bodies.  Below are 4 examples of my images taken through the Dalle 2 software.  The original images are on the left and the computer generations are on the right.

Conclusion

The sliders above show 4 different images produced photographically using oil on water to produce abstract forms. The original images appear on the left of the slider.  These images were fed to Ai using the Dalle 2 software.  The images on the left of the four sliders are Ai variations on the original image.  Ai seems to have interpreted the existing images quite well, however the final image appears to be the best.  It is strikingly similar to the original.

This technology could potentially pose a threat to artists, particularly to a photographer like myself. For example, my work could be recreated at the push of a button! The resulting image could then be used in a multitude of different ways across a plethora of platforms.  However I would be unaware it was my work being used to generate any outcomes.

This poses quite a dilemma! For example, as an artist, I can spend time creating some original artwork and then use Dalle 2 to work out a large range of variations.  In essence I have made my creation and gained a whole new portfolio from the back of one image.  However at the same time, anyone else, (without my knowledge), can ‘borrow’ my original image and again create a whole new portfolio from the back of it.  The only difference is that as the original creator, I would have been responsible for all of the hard work.  Anyone else on the other hand would not.

Therefore this throws up a number of questions like:

• Should artists use Ai to expand their portfolios?
• Should there be some form of copyright checks and balances?
• Should artists even bother creating anymore?

Technology will undoubtedly keep moving forward.  Therefore, I think it is incumbent upon us as artists, to think about ways we can creatively harness this new technology.  The key will be to make it work for us in a way that is so special it cannot be replicated at the push of a button.