Brett Welch: The Impact of Text-To-Image AI on the Creator Economy
New AI technology allows users to generate images with short written descriptions. As a result of this technology, more and more creators and brands can create high-quality images with brief written descriptions, allowing anyone with creative ideas to create stunning content regardless of artistic ability. Today, Brett Welch, Vice President of Product Management at Lightricks, shares the impact of text-to-image AI on the creator economy.
What is Text-To-Image AI, and how can creators and brands use it to build marketing campaigns and content online?
About Brett Welch
Brett Welch is originally from Australia but is living in Surrey, England. He shares that he moved to the United States when Abobe acquired his first company in 2009 and worked on the Creative Cloud. After leaving Adobe, he founded a new company specializing in collaborative video editing.
Later, he worked as a Project Manager at Facebook on video products and was even there for the ALS Ice Bucket Challenge. After leaving Facebook, he took several years off to travel before starting at Lightricks as the Vice President of Product Management. In this position, Brett manages UX designers and product managers building Photoleap.
What is Photoleap?
Brett shares, “Photoleap is a creative toolkit for photographers, artists, and designers. It’s something you can use on the go, and we have a suite of tools that lets you create anything from editing your photos, touching them up, making them look great, as well as using layers to composite lots of things together. We have a fun cutouts tool that allows you to compose and do freestyle collaging.
Motion editing tools have been recently added to Photoleap’s suite of tools. AI-Generated text-to-image features are also available. A user simply describes anything, clicks generate, and the AI creates an image based on this description in seconds. You can also sketch images within the app and add a short prompt, then let AI fill in the details and gaps in your sketch.
How Does Text-To-Image AI Generation Work?
Brett explains, “An AI model trains on a database of images, large amounts of images from all across the Internet, and it learns different concepts, so it learns what a cat is, what a dog is, and different ways that they might look as well… There’s a language model that’s being laid out over the top of this, and essentially they’re training these models to be able to generate an image from the words alone.”
For example, Brett shares that you could type the prompt “an avocado riding a horse in the desert.” The AI technology will work to imagine and bring this concept to life, creating a whimsical and fun result.
He adds, “It gets a lot more interesting or useful when you start to think about it in the context of businesses who are, for example, making product imagery. If you’re selling a coffee mug, you might want to take a photo of that coffee mug and then put it on a desk with the background set to some sort of office environment.”
This process saves businesses time and money by eliminating the need for traditional photoshoots. It also eliminates the need to take pictures or draw items. Even users who are terrible at sketching can now create unique, whimsical imagery by describing it to AI.
The text-to-image AI technolgy has been around for several years. Still, realistic images are a relatively new technology, especially with the high quality that Photoleap offers within their app.
Using AI-Generated Images in the Creator Economy
Brett shares, “The main change for creators is that the starting point can change. Rather than exploring by taking lots of photos and then seeing what works, you can start by asking AI to give you some ideas. Then, that’s a new starting point that you didn’t have before that can then spur on new ideas and boost your creativity.”
One of the best parts of AI-Generated images is that an individual’s creativity is unlimited and no longer hindered by their skill level when it comes to photography or sketching.
Brands can use text-to-image tools to generate large amounts of social media content for all their social media platforms, television, billboards, and more.
“I would say to brands, think about your best performing creative assets and think about how AI can help you by taking those assets and then generating variations on that asset that you can then potentially use in interesting and exciting ways.”
Brett shares that many brands are interested in this type of technology, but it’s a balancing act to ensure that the images generated are usable and high-quality.
Creators and brands can also collaborate using text-to-image AI. Many brands will not be able to take the AI’s output and immediately use it in marketing campaigns without polishing it first. Creators can work with brands to increase the output volume and supply great concepts for AI to work on.
Brett adds, “We’ve got our brands unit, which is called Popular Pays… it works directly with creators and influencers and puts them in touch with brands, so we’re also looking at ways that we can insert AI into this relationship to give them some sort of productivity and improvements in the process.”
A Vision for the Future
Brett explains, “The first [vision] is really empowering a whole new generation of creators… Thanks to Instagram, more and more people want to become photographers, and smartphones have enabled a whole new generation of people to call themselves photographers and learn the craft of photography.”
Brett anticipates a massive explosion of creativity around artistry and design over the next decade as more people learn how to use advanced photo-editing tools like Photoleap and AI to generate images based on their descriptions.
“I think the second thing is really around the mindset that we need to embrace, and I think this is not just creators, but in general of moving from, as an analogy, think about the advent of a calculator and how it has changed the way we do math. My father tells me about [how] he used to have books for logarithms, and you have to look up specific logarithms to find out the answer. Today, we just use a calculator.”
However, users still need to understand the basic concept to understand if they’ve typed something wrong or described something poorly.
He adds, “There’ll be a lot of creators who are fantastic at curating and editing that are going to be very successful in this next wave because they’ll be able to tell the AI what they are thinking instead of drawing it and doing work by hand.”
Creators may become more like editors or curators with this technology. For example, a creator may almost be more like a curator of an art gallery rather than the painter featured.
What’s Next for Lightricks
Currently, motion and animation features are being added and bundled within Photoleap. For example, you could take a photo from your camera roll and animate the sky or add animated stickers on top of an image, adding a unique touch to your content.
Brett shares that they are also about to release an enhancement feature that allows you to enhance old photos and even colorize black-and-white images.