Of all the industries undergoing a revolution due to artificial intelligence, the creative and design sector is one of the most vulnerable to massive disruption.
But as creative people adopt new generative AI-enabled tools that turn text into images and videos, the software companies that sell these products say the AI itself needs a bit of a redesign.
“AI promises that it'll do it all for you and magically read your mind, but in reality, you have to interact with it,” says Cameron Adams, co-founder and chief product officer at design platform Canva. “It's a bit of an art and a skill to take what's in your head and explain it, and not a lot of us are adapted to that.”
Since generative AI tools like ChatGPT started gaining traction in the second half of 2022, companies like Canva, Adobe, and Figma have been adding AI tools to boost creative productivity. Now that the industry is being asked to generate more images, text, and videos for different audience segments, geographies, and platforms like social and web, it has become easier and cheaper to design more assets.
But a lot of friction remains. Creative professionals are turning to AI to generate all of these assets, but often the tools are still clumsily integrated with the documents and presentations in which those materials appear. Lately, design software companies have been focusing on seamlessly integrating these tools across their product portfolios. “We don't look at this as a single feature,” says Noah Levin, vice president of product design at Figma. “We look at this as a technology that applies across our products in different ways.”
Australia-based Canva has been adding AI-enabled features, including a text generator that uses OpenAI's algorithms, a text-to-image converter, an AI background remover, and more. Since releasing Canva's AI-powered Visual Suite, the company says it has been focused on making the entire AI-enabled design process as consistent as possible.
“More and more people are incorporating AI into their workflows,” Adams says. Heavy investments in AI products have helped Canva add 90 million monthly active users, and its AI products have been used 5 billion times to date.
Adams said that in 2023, the generative AI boom has generated a lot of hype and niche tools for creative professionals, but many are struggling with how to incorporate these new tools into their work. AI text prompt boxes aren't particularly intuitive for users, he noted.
To make things easier, Canva is building prompts into tools like Magic Media to help users generate more accurate creative assets using AI. For example, creators can enter a prompt like “Man eating delicious pepperoni pizza” and choose from a variety of stylistic orientations that Canva offers, from dreamy to watercolor to cartoon, and aspect ratios like square, landscape, and portrait.
“Our responsibility as product developers is to make AI accessible,” Adams says. “We needed to integrate AI into our products and people's workflows in a way that made sense.”
Democratizing the creative process is important for design-focused software companies, especially as it becomes more collaborative, with concepts being brainstormed by larger groups of stakeholders rather than just designers who were previously cloistered in a studio.
“The more people working together, the better the concepts will come out,” Figma's Levin says.
Figma's strategy is to lower the minimum bar for design and give more customers the tools to be involved in the creative process, while at the same time raising the bar for professionals with a more refined offer.
“We want to attract more people into the field who have never designed before and who see great value in learning how to communicate visually,” Levin says, “but we also want the millions of people we serve, the professionals, to feel that AI can help them do their jobs better.”
No matter how advanced AI technology becomes, Adobe says humans will always need to be involved in designing creative assets. “We don't think of these models as creative,” says Adobe's chief technology officer Ellie Greenfield. “We think of them as creative assistants.”
Few industries have come under fire for negative AI-generated headlines as much as image generation tools. Google was forced to temporarily disable its image generation tool Gemini after it was found to promote racial and gender stereotypes. AI image generation tool Midjourney recently revealed that it had blocked users from creating fake political images ahead of the 2024 US presidential election.
“We go to great lengths to ensure that the images we create do not cause unintended harm or bias,” Greenfield says.
Adobe's Firefly is based entirely on licensed content, which helps limit potential bias in the company's datasets. There have been some complaints about the content generated by Firefly, but Adobe says it has established a feedback mechanism to allow users to report those concerns so they can be addressed. Since Adobe launched Firefly a year ago, users have used the generative AI tool to create more than 6.5 billion images.
As for why image generators are often at the center of ethical questions about AI, Greenfield says that when data-driven tasks are automated, humans are less inclined to lament the repetitive tasks they've handed over to machines. But because humans value art, when technology invades how art is made, it creates huge emotional problems.
“I think there's something that's tugging at different parts of our brains and our minds that's potentially inhibiting an artist's ability to create art and make a living from it,” Greenfield says.
