Hello! Welcome back to About technology: AIa pop-up newsletter that teaches you about artificial intelligence, how it works and how to use it.
In last week’s newsletter I shared the golden tips for getting the most useful answers from chatbots like ChatGPT, Bing and Bard. Now that you’re familiar with the general principle of building a relationship with AI – the more specific and detailed instructions you give, the better results you’ll get – let’s move on to a slightly different realm.
Much of the hype and fear around generative AI has been about text. But there have also been rapid and far-reaching developments in systems that can generate images. In many cases, these have a similar structure to text-based generative AI, but they can also be much weirder – and lend themselves to some very fun creative pursuits.
Image generators are trained on billions of images, enabling them to produce new creations that were once the exclusive domain of painters and other artists. Sometimes experts can’t tell the difference between AI-created images and real photos (a circumstance that has led to dangerous disinformation campaigns in addition to fun creations). And these tools are already changing the way creative professionals do their jobs.
Compared to products like ChatGPT, image generating AI tools are not as well developed. They have to jump through a few more hoops and may cost some money. But if you’re interested in learning the ropes, there’s no better time to start.
AI Photoshop
Last week, Adobe added a generative AI feature to a beta version of Photoshop, the iconic graphics software, and creators on social networks like TikTok and Instagram have been buzzing about it ever since.
I have quite a bit of experience with Photoshop. When I tested the new feature called “generative fill,” I was impressed with how quickly and skillfully the AI performed tasks that would have taken me at least an hour on my own. In less than five minutes and with just a few clicks, I was using the feature to remove objects, add objects, and swap backgrounds.
(To experiment with these tools yourself, sign up for a free trial of Adobe Creative Suite first. Then install the new Adobe Photoshop beta, which includes generative fill.)
After installing Photoshop Beta, import a photo and try these tricks:
-
To change a background, click the “object selection” icon (it has an arrow pointing to a box), then under the Select menu, click “inverse” to select the background. Then click the “generative fill” box and type a prompt – or leave it blank to let Photoshop come up with a new background concept for you.
I used these steps to edit a photo of my corgi, Max. I typed “kennel” for the prompt and clicked “generate” to replace the background. Here’s the before (left) and after.
Photo editors at The New York Times do not enhance or alter photos or generate images using artificial intelligence. But my first thought after testing Generative Fill was that photo editors who work in other contexts, such as marketing, could soon be out of work. When I shared this theory with Adobe’s chief technology officer, Ely Greenfield, he said it might make photo editing more accessible, but he was optimistic it would still need people.
“I can get really nice images with it, but honestly I’m still shooting boring images,” he said. “When I look at the content that artists create when you put this in their hands versus what I create, their material is so much more interesting because they know how to tell a story.”
I must confess that what I’ve done with generative stuffing is much less exciting than what others have posted on social media. Lorenzo Green, tweeting about AI, posted a collage of famous album covers, including Michael Jackson’s “Thriller” and Adele’s “21” which were expanded with generative padding. The results were quite entertaining.
(One note: If installing Photoshop feels daunting, a quicker way to test Adobe’s AI is to visit the Adobe Firefly website. There you can open the Generative Fill tool, upload an image, and click the ” Add” to trace around a subject, such as a dog. Then click “background” and type a prompt such as “beach”.)
More image generators
Tools like DALL-E and Midjourney can create images from scratch in seconds. They work similarly to chatbots: you type a text prompt – the more specific the better.
To write a quality prompt, start with the medium you want to mimic, followed by the topic and any additional details. For example, typing “a picture of a cat wearing a sweater in a brightly lit room” into the DALL-E prompt window generates something like this:
DALL-E, which is owned by Open AI, the creator of ChatGPT, was one of the first widely available AI image generators that was easy for humans to use. For $15 you get 115 credits; one credit can be used to generate a set of four images.
Midjourney, another popular image generator, is a work in progress, so the user experience isn’t as polished. The service costs $10 per month, and entering prompts can be a little more complicated, as you have to join a separate messaging app, Discord. Nevertheless, the project is able to create realistic, high-quality images.
To use it, join Discord and then request an invite to the Midjourney server. After joining the server, in the chat box type “/imagine” followed by a prompt. I typed “/imagine a manga cover of a corgi in a ninja turtle costume” and generated a series of convincing images:
While it’s fine to type in a basic request, some have found obscure prompts that produced exceptional results (Beebom, a tech blog, has a list of examples). At Columbia University, Lance Weiler teaches students how to use AI, including Midjourney, to create works of art.
Whichever tool you use, please note that it is up to you to use this technology responsibly. Technologists warn that image generators can increase the spread of deepfakes and disinformation. But the tools can also be used in positive and constructive ways, such as enhancing family photos and brainstorming artistic concepts.
What’s next?
Next week I’ll be sharing some tips on using AI to speed up aspects of office work, such as drafting talking points and generating presentation slides.
In case you’re wondering, the delightfully demented image at the top of this newsletter was created by a human – the illustrator Charles Desmarais – not by an AI
