With technology continuing to evolve at a rapid pace, there have been many recent discussions about whether digital renders can truly replace product photography. Going a step further, is it possible that one day artificial intelligence could simply create images without the need for any intervention from a photographer or a digital artist?
As photographers, we often marvel at how amazing modern technology can be, how magical that new “must-have” camera feature is, or how bad image processing software has become. clever. I don’t consider myself particularly old, but when I think back to using a manual focus 35mm SLR (because that’s all we had to use, not because I’m a hipster) and Whether I’m comparing this experience to the incredible face detection or eye-detection autofocus on modern mirrorless cameras, it’s hard to believe these huge technological advancements have happened in my lifetime. Even sitting in my living room, controlling home lighting and entertainment with my voice, or video calling a friend in another country on an iPad are literally things my child would have considered science fiction. Even my smartphone is significantly more powerful than any computer I had access to before I became an adult. In the grand scheme of human history, the time it took us to go from the first commercially available camera for the general public to smartphones with highly effective digital cameras in the pockets of most people in the western world, is incredibly short.
SLAB E mini
This week I saw some funny images posted on social media from a project called DALL·E mini. They were raw footage of very random things like Joe Biden eating a hot dog or spiders wearing sombreros. Silly images on the Internet are not new, but they are said to have been produced by artificial intelligence. Some of these images looked like simple drawings or cartoons, while others looked like renders from a video game in the early 2000s. None were what I would consider realistic. Most of the images shared on social media were pretty silly, so naturally I wanted to research where those images came from. A short Google search brought me to DALL·E mini. This Transformer-based text-to-image generation template was designed by Boris Dayma, Suraj Patil, Pedro Cuenca, Khalid Saifullah, Tanishq Abraham, Phúc Lê, Luke, Luke Melas, and Ritobrata Ghosh.
DALL·E mini is very simple. You type a short text prompt, then the AI, which was trained on unfiltered data from the internet, goes to work and produces nine images based on the text stimulus. These images usually vary a lot from each other, but represent the AI’s interpretation of your input, based on data from the internet. Right now it’s not particularly fast, taking between two and four minutes to produce images that are questionable quality at best. After playing way longer than I should have, I can see that it’s nothing more than a gold mine for memes right now, but as a concept it’s fascinating, with exciting future possibilities.
The template is intended to be used to generate images based on text prompts for research and personal consumption. Intended uses include supporting creativity, creating humorous content, and creating generations for people curious about the model’s behavior. Intended uses exclude those described in the Misuse and Out of Scope section.
It should be noted that these images are created by an artificial intelligence, which has been trained on unfiltered data found on the Internet, to produce its own interpretation of the search terms that users give it in the form of a selection of basic pictures. It should also be noted that netizens use their own creativity and imagination to instruct this AI to create comic effect things. If you are considering checking out the discussion forum or trying out the image generator yourself, be aware of the biases and limitations text provided by the development team, and be aware that some people on the internet are jerks who will find it funny when the AI produces dubious or offensive images.
While the capabilities of image-generating models are impressive, they can also reinforce or exacerbate societal biases. Although the extent and nature of biases in the DALL E mini-model have not yet been fully documented, since the model was trained on unfiltered data from the Internet, it can generate images containing stereotypes against minority groups. Work to analyze the nature and extent of these limitations is in progress.
The future of AI-generated images
It’s probably safe to say that no photographer will soon lose their jobs because of AI. This technology, however, raises some questions about what the future of imaging might look like. We now live in a world where stock images are available online in seconds for anyone who needs a generic image. Sure, the stock images were taken by a creative professional who will earn an income from them, but what will happen when machine learning gets to the point where some generic images can be created by the ‘AI? Who owns the rights to these images? Could this one day replace much of the stock image industry and hurt stock library photographers? Could we one day see renderings of products or places entirely produced by a machine algorithm being used for commercial purposes?
Memes and silly images aside, I wanted to see how close this system was to creating a realistic landscape, so I gave DALL E mini a simple text input to see what it did. would make a beautiful landscape as a text input. Here is the image it produced this morning. Attention landscape photographers! The machines are coming for your work!
I appreciate that from the looks of the images being produced today it seems like a stretch to think it could ever replace a professional photographer, but 30 years ago an iPad and FaceTime were science fiction, but now we all carry tiny powerful computers with megapixel digital cameras in our pockets every day. The possibilities for the future are exciting or terrifying, depending on your perspective.
Renderings, which are still created by humans, take over from product photography in some places. Is it only a matter of time before digital images are so lifelike that we won’t need real photographers in so many situations? Is it possible that there will be no need for commercial photography at all one day?
What do you think of renderings or AI replacing photography? Is this technology exciting or worrying? How far could a legitimate commercial use of this technology lie? Let me know in the comments.