In less than a year, generative AI has evolved faster than I ever expected—not just the technology but its mainstream awareness and popularity. Generative AI-powered chatbots are conversation partners and writing coaches, while image generators allow anyone, regardless of experience or skill level, to express their creativity in ways they never imagined. It's a wild time to be a professional creative, and as a director of design at Adobe, whose team is helping develop tools that will define one of the most pivotal times in the history of creativity, it's been even wilder.
In March, we introduced our own creative generative AI model, Adobe Firefly, and it far exceeded our expectations, with users generating more than 100 million images in just the first six weeks. The enthusiastic reception from the creative community wasn't an accident. It was the result of very deliberate thinking, planning, and innovation across the product design process and deep involvement from the creative community.
Putting creators at the center
As both designers and users of Adobe's tools, my team has a unique opportunity to speak with conviction and advocate boldly for creatives to design generative AI tools the way we need them to be – or else others will make the decisions for us. As the design lead for Firefly, I know that doing so is what's helped make our solution so special.
The first time I walked into an Adobe office, I wasn't an employee; I was a customer: A freelance designer with a degree in fine art and a background in drawing, painting, and graphic design, who used Adobe's products to make my living. Many members of the Adobe Design team are also artists, photographers, and videographers.
We wanted to think differently about the generative AI space, so Adobe Design formed a new hub: The Machine Intelligence and New Technologies (MINT) team, a group of designers serving as the center of gravity developing a design-led strategy for emerging trends and new and advanced technology. This team intentionally designs generative AI tools from a holistic perspective, talking to customers, bringing their own experiences to the table, and influencing every aspect of the experience. We're involved not only in designing the UI and UX but also in how the technology is built and positioned within users' creative workflows.
With Firefly, we knew the right design could help accelerate the creative process, not replace it. We imagined that more productive and accessible creative tools for millions of our customers around the world would have an important and positive impact on the future of creativity for creative professionals—and for everyone else.
Less pain, more fun and focus
Our goal is to help creatives spend more time being creative and less time dealing with the hidden production work that robs us of the joy of making. When thoughtfully designed, generative AI can dramatically speed up work, from inspiring ideation to removing tedious and time-consuming tasks. In addition to working with photographers, graphic designers, and illustrators, our design team sought out specialised creative professionals, including comic book and storyboard artists—people who helped us uncover and address particularly painful friction points across multiple industries. Removing as many roadblocks as possible provided a quicker path from an initial moment of inspiration to having something beautiful on a canvas, and it freed creative people to focus on the work they enjoy.
Designing a more human interface for generative AI
Available through chat interfaces, catering to limited communities, and often outside of the creative mainstream, most generative AI interfaces are not very accessible. For most people, composing a prompt and getting compelling results from generative AI feels a bit like a secret handshake.
When Adobe Design's MINT design team started talking to creatives and watched them interact with the technology, we quickly realised what needed to change. Digging deep into my own experience as a user of Adobe tools, I spent days going down rabbit holes, researching prompting language, and experimenting; it took a lot of time, and there was a lot of friction involved in the process of getting useful or interesting results.
The first time you get something you love from generative AI, it's fun! Successful creation is incredibly gratifying, and we wanted people to get that feeling with more speed and less friction. Rather than providing a blank canvas, we built a scrollable gallery of sample images that, on hover, reveals the text prompts along with a friendly "try prompt" button to kickstart ideas and encourage experimentation. We also partnered closely with our engineering team, sharing datasets and wading through the nebulous discovery process inherent in artistic creation. As a result of this artistic and technical collaboration, we were able to add a range of visual styles, including Baroque, Cubism, and Cyberpunk, each of which can be applied instantly to any generated image.
Building in trust
It's common for creative professionals to work in settings where a large part of their job is running a responsible business and interacting with clients. While they need tools that help them work faster and produce brilliant art, they also need tools they can trust in commercial settings. The MINT team is thinking about every aspect of the experience as part of the design process.
Like all Adobe AI innovations, Firefly was developed according to our AI Ethics principles of transparency, responsibility, and accountability. Firefly is also trained on a unique dataset that generates professional-quality content and is designed to be safe for commercial use. And through the Content Authenticity Initiative, which Adobe founded and now has more than 1,000 members, we'll attach digital "nutrition labels" that provide transparency on how the content was created, so creative professionals can feel confident in using this transformational technology in their professional work.
Broadening the Aperture
Image generation was an important first step, but our design and engineering teams—and creative community—wanted to take generative AI further. Graphic designers told us they saw potential in text effects; illustrators said they wanted text-to-vector capabilities; and videographers hoped text prompts could help them create, modify, or edit video. We continually listen to community input and use feedback to shape the future. Today, all those capabilities are either already live in the Firefly beta or are in active exploration.
Firefly has launched, but we're not done building. Our work on generative AI is always evolving, and we'll keep listening and integrating feedback from our customers and the creative community as their needs, the technology, and the policy landscape evolve. If you're a creative pro or anyone with a story to tell, we invite you to try Firefly and take part in designing the future of creativity.
