Artificial Intelligence has been a big topic this year in both positive and controversial ways, primarily for its advanced writing ability in most AI models available today. With this, strikes have already occurred, jobs have been put at risk and cheating within schools has risen. AI has been around and has made its debut since the start of this year with ChatGPT, one of the most popular AIs, reaching its first birthday this November. Writing has usually been the only subject amongst AI, the second most being its ability to generate images and what can be called art.
Unlike AIs’ average ability of text generation, images are not a strong point. There are many different websites and software that provide this image generation service. But generating a picture is so much more than writing text based on a prompt, leading to a weakness of sorts. AI-generated images have already been proven to either work or fail, producing something of quality or a shriveled-up mess. It all depends on what one desires to be generated.
But just like its ability to write, AI is constantly learning, improving, and only becoming more and more capable of doing what it struggles to do. One can only keep learning for it to get to a point of “intelligence” where it keeps decreasing its limits. Money of course has a big factor in it as well, as smarter and more advanced AIs require a subscription to the software. The price and access all depend on its overall ability.
One of the recent paid art and image AIs was provided by Adobe, implemented into a beta version of Photoshop. At first glance, nothing is too different about this version. Though a right click will reveal a brand new photo editing tool: generate fill. Highlighting anything in an image allows it to be replaced by a prompt, creating an AI generated image to replace the section, to a point of blending in with the image’s environment too.
The Photoshop AI, as most others, is still not as powerful as it could be. The biggest challenge is making it generate what you want most of the time. No matter how specific you get with your prompt, there is a good chance that you will not be provided with what you want. The generation process takes up to a minute and makes four images at once. Selecting one will replace the highlighted section with the image, fitting in as much as possible.
The biggest difference between Photoshop and other image AI is its ability to fit art and pictures into an image the user selects. This is the reason I enjoy it so much. This is mainly because I like to see how it handles my prompt while seeing how well it blends the image into my image. It is something new and different implemented into the “world of AI,” with other producers likely to follow with similar features and capabilities.
Even though the world of AI is expanding more every day, it is cool to see the developing capabilities, while AI is not very advanced yet. Just like the writing situation, the same could easily happen in the art and photography industry. As of right now, let us try and enjoy this strange software before the bigger picture becomes possible to bite us.
