Artificial intelligence has recently entered the art world in a big way.Companies like Nvidia have been working on AI tools to make art with basic curves (opens in new tab)And robotic AI artists are doing unique and interesting paintings. (opens in new tab) Recently, the internet has been flooded with images created by the DALL-E 2 art robot, which can create images with just a few words. They’re often the perfect mix of impressive and utterly damned, like these AI-generated Duke Nukem (opens in new tab).
While these AI art bots are able to generate fresh meme content on a regular basis, they are their core, tool and a photographer (by PetaPixel) found the software to be of great use. Instead of creating the entire image as described in DALL-E 2, they used it to clean up the shot, and the result looked more impressive than Photoshop.
Nicholas Sherlock (opens in new tab) is a photographer living in New Zealand who used the DALL-E 2 to turn some very nice but blurry photos of ladybugs into sharp stock photography style images. To do this, he had the AI edit his original blurry photo and display the text prompt “Ladybug on a leaf, focus stacking high-resolution macro photo.”
Sherlock then showed these images to photography YouTuber Micael Widell (opens in new tab) They show them off in a video exploring the use of DALLE-2 in photography and potential murder. This video starts with a few different examples of the DALLE-2 before showing off Sherlock’s example in photo editing, all with results that are at least somewhat impressive.
The video also shows another example where DALLE-2 is asked to generate other variants of the provided image, and the results appear to be indistinguishable. The capabilities of these AIs for image generation are impressive. PetaPixel also compared more images provided by Sherlock. When compared to the restored image produced by Photoshop, the DALLE-2 clearly wins.
The ability to perfectly edit an actual photo the way you want does seem like a good use for this tool, especially in a result like this. To be fair, Widell wonders in his video what the role of photographers and even hobbyists is in the future of DALLE-2 AI, where anyone can generate any photorealistic image they want. Of course, there will always be other reasons to pursue these passions, but especially when it comes to things like stock photography, these AIs are sure to feel like a game changer.
We’re likely to start seeing more and more AI-generated art being used in games, whether to come up with concepts and explore options, or even make usable assets. There are many questions about what this means for the future of visual arts in many industries and hobbies. However, it is important to never underestimate the human desire and joy in creation. Hopefully, these AIs can be a tool to do this more often, rather than take it away.