Funny Things I've Generated Using Generative Fill in Photoshop

After delving into generative AI tool in Photoshop, I decided to put Generative Fill to the test and generate some funny and some nightmarish solutions of my own. Below you will see some of the outcomes. Generative fill creates 3 solutions, I have chosen the most fitting to post here. Generative fill as you will see, can be scarily accurate and other times just hilariously wrong it its outcome. It's not quite there yet in terms of accuracy, but it won't be long until an update comes along making it so. 

A happy ferret

This solution is rather fitting. I don't have any issues with the image above, the ferret is smiling though unrealistic. I did not give the AI any other words/directions, therefore its done a decent job. 

A dog riding a pink scooter through traffic on a rainy day

As you will see, the majority of this list is animal themed, to keep things light and humorous. They also seem to be the most fun to generate, being their limbs work differently to ours, so it is fun to see how the AI presents these doing human things. Though you will see above, in this example it has done a stellar job. Though I would argue that the dog isn't riding a scooter, but a hoverboard. Nonetheless, shown this picture to someone years ago they would have possibly believed this was real. 

Cat conducting a choir

Interestingly, the AI has decided to dress up a cat in a formal blazer. I suppose this goes with the formal attire you may traditionally see on a conductor. Though it is just a cat, wherever the AI bases its images on will always deem to be biased toward a certain class/way of dressing etc, as it will be based off whatever it has. Therefore wherever its information for conductors comes from, they are classy!

Capybara flying a plane

In the summer I had an obsession with Capybaras, therefore I thought it was only right to generate some capybaras doing things. Seeing as I had just gone away, I thought to generate it on a plane or flight. No solutions showed the capybara on a plane, but they were all plane-beings instead. This was kind of cool to see and the wings work really well. They are not plane wings but animal wings. Perhaps we have just spawned a digital species of flying capybara.

Eerie colourful city skyline

The game City Skylines was my inspiration for this one. Often times colourful can seem counterintuitive when talking about the scary and eerie, so I wanted to try the two together. Photoshop has done a fantastic job and morphing an eerie looming skyline with a nice colourful sunset. I have seen digital artists who Illustrate in a similar way. Therefore I think it's always important to fact check the artist and their workflow, as I would have possibly fallen for this on Instagram. 

Capybara riding a bicycle

Another episode of "Capybaras Doing Things." This time I generated a Capybara riding a bike. This was the most realistic one, however notice there are no pedals etc, though their feet are in the right places. The capybara has been propped on the bike like a saddle. It's not a good generated solution but definitely one to show. 

Shrimp Chimpanzee (Shrimpanzee)

This is an illustration idea I had had in my head for a longtime, being shrimp and chimp rhyme. What Photoshop has generated is the exact opposite of what I would be going for... THIS IS NIGHTMARE FUEL. Why does it have a hard shrimp shell with chimp eyes? It's quite threatening and hard to look at. 

Cute little spider sitting in the sun reading a book cartoon

I wanted to try adding a style to see how the fill would interpret it. The illustration it has created is really great in terms of speed, however I wouldn't say this was the kind of cute I was going for. There may be someone out there who would find this fits the brief, though for me it's far from what I wanted. The day isn't sunny and quite frankly its eyes are a little garish. 

One of the many things that separates Illustrators apart from AI is during the process of onboarding a client, they can pick up on cues other than words etc that can help finding out what a client is looking for. AI can only pick up on words, therefore Art and Design workers will always been one ahead of an AI. The service is tailored. Clients often cannot put into words what they want, it is the workers who show and tell their abilities and eventually come to a final solution. The industry is going to be just fine. 

Though I do find the new abilities to do this fun and possibilities endless... I don't think it'll take over just yet.

How I Feel About Photoshop's Generative Fill

Generative Fill in Photoshop has been a widely discussed topic, as well as AI in general. If you are not sure what Generative Fill is, it is a new tool in Photoshop to which you can select an area, and instead of Content Aware fill etc, you can opt for Generative Fill, to which you can type words to instruct Photoshop on how you would like it filled. 

(There is also the option to leave it blank and let Photoshop figure out for itself. However so far, I find the outcomes of this way to be quite garish)

Within design, this has many advantages and disadvantages. I have heard first hand that many Agencies are beginning to offer the use of AI within their client's outcomes. If it is going to be both widely-used and widely available, why shy away from it? A good Agency and Designer understandable needs to stay current and change with the industry, in order to offer the most cutting-edge and modern service for customers. 

I have been struggling to use Generative Fill. If I have been able to create anything I deem usable, it is still evident as of the date this post has been written that it is not quite right. Some of the generated images are not accurate in terms of both size, realism and the fit to the selection surroundings. It is particularly obvious when generating faces or parts of faces. Though, through research I am aware there are other AI programmes/software that can achieve real likeness and this is scary. 

However, one has to wonder where the inspiration for the AI's outcomes is coming from, and how will this look for the future of copyright. These were my two greatest concerns. I also am curious as to how this will also look for the future of the work involved in Graphic Design, and whether this whole process in the future will become at some point automated.

I began to collate a list of advantages and disadvantages, however I realised as I was writing how conflicted I was on the topic and that some of the points I was making had a great counterpoint. Therefore see below my list of arguments relating to Generative Fill.

Arguments surrounding AI and Generative Fill

Ironically, this will make some aspects of design more accessible for designers. It sounds counterintuitive I know, however, small details such as adding a bow around an object or an object in the background can be make or break for a poster or image and is cheap and easy to businesses who cannot afford good mockups and back drops. 

It will make some tasks in Photoshop quicker no longer needed. Generative Fill may be able to complete some tasks, such as adding to an image or correcting parts of an image. This could be good for designers with approaching deadlines, or needing to do something in a rush. Over the years of using Adobe, they have added many updates to their programmes which have made some of the tasks I once completed quicker and easier with a new way, however I find I am most comfortable still staying to the old way I used to do it. Humans are creatures of habit, and we don't like change. I am evidence of this. 

During the age to which I have entered Graphic Design, using technology and Adobe software being known as the 'industry standard' has been the norm. I have not known Graphic Design without it, however I great respect for those who can or have used the hands for design in the past or present. Considering the manual work that goes into such craftsmanship, I think the respect for design completed without the use of AI will only increase tenfold. 

Those new to Photoshop could become accustomed to using Generative Fill and perhaps not learn the manual ways of working. This is a fear of some I know, however if it achieves the final desired outcome should we really complain? It's hard to see this happening to our beloved software packages, although is it a resistance to change and working smarter not harder? I am very conflicted around this point. 

A summary

So how do I feel? Alike other designers, I feel conflicted. 

I am for positive change with design and making ways of working easier, however going back to my initial concerns I am scared about originality within design. Will AI be able to achieve such personalised and unique outcomes as unique as the designer who could have created them? 

This is a current and changing topic so I will be all-ears for new updates and the future uses to be seen of AI. I aware some countries/laws will be banning the use of AI in certain contexts. I am glad government bodies and those in industry are reacting to this in such as way, as it is a great change to what has been known for so long. 

I could talk about this topic for a long long time and the greater topic of technology and its uses as a whole, how it harms and benefits us. However I would not have enough time nor energy. 

I am struggling to finish the end of this post because I am so conflicted on it. I personally will not be using it due to its credibility and also how much I realised I actually enjoy about the manual tasks within Graphic Design. There is a sense of earning something / owning something when you know you have bought it from conception to its final outcome all on your own. Though, I do see its uses.