AI Image Generation
A few days ago a photographer friend, Steve Peixotto, suggested I give the AI image generator MidJorney a try. I was floored. I still am floored. You no longer need to be talented to create amazing art. Just type in some “prompts” and Voila! Art! Sometimes really crappy art, but sometimes really amazing art. I remember back in the mid 90’s I played around with some of the first digital cameras and remember thinking that while useful in some ways digital capture would never be as good as film. Boy was I wrong. Now days film can’t hold a candle to digital capture. Someday, probably sooner rather than later, AI image generation will be superior in just about every way. One question I have is whether the negative will out weight the positive!
An AI Sample Image…And Copyright Concerns
Steve sent me an image he created with Midjourney by prompting “Old woman with neck tattoo”. I am not sure if that was his exact prompt…but I followed up by prompting “old woman with neck tattoo” and got a very similar image. Both images can be seen above. On the image from my prompt the woman appears to have a large open sore on her nose…nice touch! I experimented with about twenty or so images until I had used up my free allotment. The results were somewhat inconsistent, but overall I am staggered by how good the results were and with all the possibilities presented by this technology. One big question is how copyright issues will be resolved. Some agencies such as Getty are working on AI image generators that will theoretically solve that problem by working with images within their collections. But MidJourney, for example, is “Trained” using images, according to their own CEO, from the Internet. In theory there are probably millions of copyright infringements in that process alone…and I heard that Getty, for one, is suing at least one or two AI companies for that very reason. I see a huge copyright mess out there in regards to AI image generation!
Stock Agency and Freelance AI Artists
In thinking about how the advent of AI will change the photography business, I can see how an art director might go to an “AI Artist” to fill his or her creative vision. The AI artist will be an expert at what prompts are needed to achieve the art director’s (or other client) vision. The AI Artist will know all the variables to efficiently achieve a desired result. I can see a world where stock agencies have in house AI artists who use the agency’s image collection to solve copyright issues. Even as we speak (okay…type) Getty is partnering with an AI company to that end. Undoubtedly there will be freelance AI artists who can perhaps utilize various image databases. While I haven’t researched all the different AI generators available, I do know the Shutterstock has an AI generator as part of it’s offerings, and Adobe is in beta with it’s Firefly image generator. Depositphotos offers an AI-generated image service called “Crello” that allows users to create custom graphics and designs using a library of AI-generated templates. Canva offers an AI-powered design platform that allows users to create custom graphics and designs using a library of AI-generated templates and design elements.
AI Will Disrupt The Photography Industry
The royalty free model of stock photography was initially fueled by art directors and designers who were doing there own photography and sharing it online. I can see the same thing happening in the AI image business as well. Perhaps some entirely new form of “stock agency” will result. Guessing the future is hard…but there is no doubt in my mind that AI image generation is going to disrupt the photography industry…and it may well be better for photographers to be at the forefront than under the rear wheels of such developments. But overall, I am afraid it is going to devalue photography in general.
Photoshop And AI
I can also see a world where AI image generation becomes just another tool for photographers. They can use AI to create backgrounds or props and can use Photoshop to combine real and AI elements, or just different AI elements. Perhaps it is the photographer’s vision that is sought and it is up to the photographer whether to use AI or traditional tools to fulfill an art director’s requests. I am sure that AI generated imagery will often be refined and “finished” with Photoshop.
Fine Art Photography And AI Generated Images
With the maturation of AI image generation the vision will become more important than the process. I used to say that Photoshop eliminated the barriers between imagination and execution. That becomes exponentially more true with AI. Of course, amazing visuals may well become a commodity…and then it will be marketing and branding that determines the value of artistic endeavors. Actually, that might already be true!
Big Problems With AI Image Generation
AI image generation scares me. As a disinformation tool it has tremendous power. Just it’s existence is a threat because the knowledge of its existence undermines our confidence in the information we take in. It used to be that we generally believed that what we read was true, and that photographs don’t lie. But now what we read is often untrue and photographs can lie very effectively. While photography has always been capable of lying through selective cropping, and more recently Photoshop sleight-of-hand, one always had to have the raw material to work with, and the skills. With AI image generation you just need to type in the prompt. What happens to our society when one is forced to doubt all information sources? Or, perhaps, when one doesn’t doubt them enough? Neither option is good. I forsee ever more polarization and chaos as deep fakes begin to multiply as I am sure they will…and are already doing. The prevalence of deep fake imagery can erode trust in media and other forms of communication. If people can’t trust what they see and hear, it can be challenging to make informed decisions and engage in meaningful conversations. Deep fakes can be used to manipulate financial markets, spread false information about products or services, or even to undermine national security. Imagine, deep fakes can be used to create fake news stories, political propaganda, or even pornography. They can be used to impersonate individuals or to create false evidence in legal cases. In addition, deep fakes can be used for cyberbullying, revenge porn, or other forms of online harassment. While deep images have always been around, and much more possible with Photoshop, AI image generation democratises the process making it increasingly easy and universal to produce. It is truly scary to think about all the damage that can be created to individuals, companies or countries by AI image generated content. All you need now is a smartphone and Internet connection to potentially wreak untold havoc on the world!
Dealing With AI Image Generation Risks
To mitigate the risks posed by deep fakes, it is important to develop better detection and verification techniques to identify fake images and videos. It is also important to raise awareness about the risks of deep fakes and to educate the public about how to spot them. Finally, there is a need for policymakers to regulate the use of deep fakes and to ensure that they are used in a responsible and ethical manner. At this point the tech companies have generally failed to adequately address the problems associated with disinformation. And it is only going to get worse. Tech companies must also step up to plate in dealing with the problems of AI image generation.
What AI Has To Say About AI Image Generation
In the spirit of all things AI, I asked chatgpt to comment on the impact of AI image generation on the photography industry and on society. Here is what the reply was:
The advancement of AI image generation technologies is likely to have a significant impact on the photography industry and society as a whole. Here are some potential effects to consider:
Disruption of the Photography Industry: As AI image generation technologies become more advanced, they may be able to create images that are nearly indistinguishable from those captured by a human photographer. This could lead to a decline in demand for traditional photography services and a shift towards AI-generated images.
Increased Accessibility: AI-generated images could make it easier for people to create high-quality images without needing expensive equipment or technical expertise. This could democratize the field of photography and make it more accessible to a wider range of people.
Ethical Concerns: There are ethical concerns associated with the use of AI-generated images, particularly in areas like journalism and advertising. If AI-generated images are used to deceive people, it could erode trust in the media and have negative consequences for society as a whole.
Creative Possibilities: AI-generated images could also open up new creative possibilities for photographers and artists. They could use these technologies to experiment with new styles and techniques, or to create images that would be impossible to capture with traditional photography methods.
Impact on Employment: The rise of AI-generated images could have an impact on employment in the photography industry. As demand for traditional photography services declines, some photographers may need to adapt their skills or find new ways to use AI technologies in their work.
Overall, the effects of AI image generation on the photography industry and society are complex and multifaceted. While there are certainly potential benefits to these technologies, there are also significant ethical and social implications that need to be carefully considered.
