https://static01.nyt.com/images/2023/06/02/business/00ontech-ai-nl/00ontech-ai-nl-facebookJumbo.png
Welcome to On Tech: AI, a document intended to teach you about artificial intelligence; how it works and how to use it.
In the last post, we provided tips on how to get the most helpful feedback from chatbots such as ChatGPT, Bing, and Bard. Now that you know the basics of building a relationship with AI – the more specific and detailed advice you give, the better results you’ll get – let’s move on to the next area.
Artificial intelligence has caused fear and panic in the literature, but there have also been rapid and dramatic developments in imaging systems. For the most part, these systems share similar features to text-based AIs, but they can also be very weird – and lend themselves to interesting creative possibilities.
Image generators are trained on billions of images, enabling them to create new designs that were formerly only made possible by artists and other experts. Sometimes experts cannot even differentiate between an AI-generated image and a real one (a situation that has inspired a well-known negative campaign, in addition to interesting creations). These tools are already changing the way design professionals work.
Compared to things like ChatGPT, AI graphics tools are not as advanced. They require users to jump through additional hoops and spend a little more money. But if you want to learn the ropes, there’s no better time to start.
AI Photoshop
Last week, Adobe added an AI feature called “generative fill” to the beta version of its photo editing software, Photoshop. Creators on social networks such as TikTok and Instagram have been talking about it ever since.
As someone experienced with Photoshop, when I tested the new feature I was impressed by how quickly and efficiently the AI performed tasks that would have taken me an hour to do myself. In less than five minutes and just a few clicks, I was able to remove and add items and change backgrounds using the generative fill feature.
(To try these tools for yourself, start by signing up for a free trial of Adobe Creative Suite. Then install the latest beta version of Adobe Photoshop, which includes more features.)
When you have Photoshop beta, upload an image and try these tricks:
-
Changing the background: click the “select item” icon (an arrow pointing in a box), then under the Select menu, click “adjust” to select a background. Then click the “generative fill” box to quickly fill it, or leave it blank to let Photoshop bring you a new background concept.
I used this to edit a photo of my corgi, Max. I typed “kennel” to tell it, and clicked “create” to change the background.
Image editors at The New York Times do not modify or enhance images, or create images using artificial intelligence. But when I tried generative filling, my first thought was that graphic designers who work in other fields, such as advertising, may soon disappear. When I shared this with Adobe’s chief technology officer, Ely Greenfield, he said that it would make image editing easier, but he was optimistic that people would still need it.
“I can make really beautiful pictures with it, but to be honest, I still make really bad pictures,” he said. “When I look at what artists make when you put this in their hands versus what I make, their stuff is more interesting because they know how to tell a story.”
I admit that my experience with generative fill is less interesting than what others have been posting on social media. For example, Lorenzo Green, who writes about AI, created a collage of famous album covers including Michael Jackson’s “Thriller” and Adele’s “21,” which were boosted by a complete background change. The results were very captivating.
(One
