From Pushing Pixels to Propelling Creativity
How Creative Leaders Will Empower Teams Through AI
“A computer is a bicycle for the mind.”
— Steve Jobs
Creating and sharing original ideas, words, art, and products are intrinsically linked to our ability to craft and utilize tools. The extension and amplification of our capabilities through technology is a key advantage that humans have leveraged to accelerate our progress throughout history. Tools of all kinds have helped to scale our capabilities, allowing us to create things with consistency, quality, and volume. Over time, more and more people gain access to evolving tools and processes and benefit from the ever-increasing advantages they provide. Our fabrics were once woven by hand. The development of looms allowed for the broad availability of consistent, high-quality, lower-cost textiles. We once hand-crafted wooden carriages and attached them to horses for animal-powered transportation. Then we built steel machines powered by engines, and produced them in massive quantities, allowing more and more people to gain access to rapid transportation.
Digital technology has accelerated automation at an exponentially greater pace than in prior technological eras. The first generations of low-fidelity digital images were created by individuals with specialized knowledge of computer programming. Now, the moment I touch the tip of my Apple Pencil to the screen of my iPad to illuminate a single bright pixel of light, I am leveraging a highly automated process. My analog tactile input is automatically translated to a digital medium – without requiring me to write a single line of code.
But we have a complicated relationship with technology and automation. Even though we understand that advancement is beneficial to our society, individuals are often resistant or feel threatened by advancing technologies. This is because automation sometimes displaces craftspeople who have spent their lives in a particular area that is being encroached upon by new machines, capabilities, and processes.
Designers and artists have traditionally relied upon some degree of highly trained, hands-on human interaction to create. Painters use a paintbrush. Writers type or dictate. Even the relatively recent advent of computer-generated imagery requires manual work and direct manipulation by artists and technologists. But now, artificial intelligence and machine learning are beginning to introduce significant advancements for automation within the creative fields. It is early, but there is incredible potential for creatives to pioneer in this new space, as well as significant challenges for leaders as they help their teams build new skills and perspectives as they navigate through this frontier.
If, as Steve Jobs once said, a computer is a bicycle for the mind – accelerating and amplifying our intellectual capabilities – then creative applications of artificial intelligence just might be analogous to an electric car for inspiration.
What is it and how does it work?
There are many areas where AI is pushing the boundaries of creativity, including writing, conversational design, and even computer programming. But let’s focus on a specific area where AI is introducing bold new capabilities today – generative visual art.
Generative AI art is produced through a few fundamental processes and capabilities. The first aspect is the leveraging of machine learning to train a computational neural network. Digital images and photographs are defined and indexed by text-based descriptions for each image. Throughout this process, the system is taught to identify specific objects, differentiate between objects, states of objects, as well as determine relationships between objects. Combine this knowledge with the ability to compose, iterate, and render images with different visual styles, and you start to get the picture.
Several different experimental systems are being developed today, including Dall•e and Midjourney. An individual typically interfaces with these systems through text-based input that includes some basic parameters and keywords for what you want to visualize. For example, if I wanted to create an image depicting a future where the Mayan civilization has achieved space travel, I could type something like this into the Midjourney interface;
“/imagine futuristic people walking through a mayan spaceport”
Midjourney has produced four pieces of art based on my direction. Each one is slightly different, and you can see that the system reflects styles, colors, and motifs vaguely resembling those of ancient Mayan architecture and artifacts. It also interpreted my request for “people” to be included in the scene as two central people, not a dozen individuals, or even three, which is also interesting. I like the results, but I’m going for something a little more moody.
“/imagine futuristic people walking through a mayan spaceport at night concept art” renders four new images.
Now we’re getting somewhere! The simple addition of a time of day has created a darker image, as well as the inclusion of the term "concept art" has resulted in a more stylized direction. At this point, I have some options for how to further develop this piece. I like the direction of the lower left image, so I asked Midjourney to focus on that image and produce four new iterations based on that specific asset.
From here, I like the atmosphere and detail of the lower left image, so I ask Midjourney to render a high-resolution version of that specific image as my final piece of work.
The result is pretty amazing. But not without some shortcomings and limitations. Each iteration is slightly different in terms of the specific details it is rendering. Look at the differences in the shapes of the pyramids and lighting between the thumbnail and the final asset. The level of control in terms of tuning is limited. While you can add a variety of known artistic styles (to create a pastiche of Van Gogh, Picasso, Dali, Hopper, Warhol, etc.), the system only knows what it has been taught, and the interpretation of your prompts varies widely. There is also a simultaneously mesmerizing and frustrating level of “dreaminess” infused throughout the pictures. Forms are often recognizable but somewhat distorted, often leading to a surrealistic undertone. That said, working with this system is amazingly fun and addictive and before you know it, you’ve been creating things previously unimaginable imagery for hours on end.
Cool for dreamy, warpy, painterly pieces, but how about more graphic elements? If we were looking to create a logo for this newsletter, using the letter “C”, you would see some iterations like these:
One of the most amazing things about this technology is the ability to synthesize ideas to create compositions based on different and previously unrelated concepts. What if Leonardo da Vinci designed an electric car? His sketches might look something like this:
The impact of AI on the creative arts is early, but rapidly gaining steam. Popular media is not only directly leveraging these new capabilities, but also incorporating the process of creating with AI-based tools into their storytelling. One example of this is the character Christina (portrayed by Evan Rachel Wood) who is a story writer in season 4 of Westworld. The tool she uses to conduct her work is a 3D holographic workstation. She speaks to the system to compose scenarios for interactive game experiences. The workstation interprets her verbal direction to render real-time animated iterations of her pitches. We are a long way off from this technological vision. But not as far off as one might think.
Three Approaches to Empower Creative Teams Through AI
Helping teams successfully navigate through change is an essential responsibility of any leader. Leaders of creative teams have two added elements of complexity. The first is to challenge their teams to evolve ahead of future consumer and industry expectations, while simultaneously providing uninterrupted delivery of current priorities and commitments for their enterprise. Looking ahead but delivering for today is a real - but vital - juggling act. Second, introducing new AI-powered tools and techniques to teams is a long game. While still largely experimental today, these tools and processes are rapidly emerging capabilities that will become increasingly important in creative endeavors.
Here are a few considerations to include in your leadership strategy:
Focus on premise, not pixels – As the title of this newsletter suggests, some aspects of “the work” will focus less on manual input and manipulation to deliver creative output. Instead of pushing pixels around on a screen, connecting vertices, filling spaces with color, or even typing every word in a movie script, artists, designers, writers, and content makers will begin to operate more like creative directors. Instead of directing teams of individuals, they will direct capabilities made accessible through AI. Creators will formulate a vision for where they want to go, and then direct a system to produce a result. Because AI systems learn and improve over time, this process will be iterative and fluid, producing not only singular results but many options. This is a huge advantage for creatives because they will be able to generate more ideas at a rapid pace than ever before.
The mechanics and dexterity that visual artists currently rely upon may become less important in the future. AI-powered tools will allow more people to express themselves creatively than ever before, with little or no formal artistic or craft-related training required to produce highly detailed assets. Exceptional creative output starts with a foundational knowledge of the craft. Understanding graphic standards, learning about aesthetics, usability, composition, typography, color theory, storytelling, editing, and so on will be required to maximize the potential introduced by these powerful new tools, and results that resonate with audiences and consumers.Develop creative strategists – Digital experience designers, who design apps, websites, and other digital services, operate at the epicenter of the development process. To maximize their impact they build connections between many different disciplines, including product management and software engineering, listening to customers, interpreting requirements, introducing innovation, and applying consistency, to ultimately bring organizational capabilities together to produce a compelling product experience. This is a unique position to occupy for individuals who apply a longer-term, strategic approach when it comes to the introduction of new ideas, processes, and capabilities.
Successful creative strategists are curious. They are curious about not only current consumer expectations and enterprise-level goals but about what might be anticipated and required in the future. Curiosity, combined with imagination and insights to inform what might be possible in the future, illuminates the future for any endeavor. Help your teams connect the dots between established best practices today, and what will be possible tomorrow by applying emerging capabilities to your work as soon as possible. Even applying AI-generated art to incubator programs or experimental tests will allow your teams stay ahead and grow through real-world experience with new capabilities.
Leverage AI as an assistant – AI tools are still very experimental and have significant limitations. The systems are only as good as what they have been taught. Once you start to play with these systems, you can pretty quickly find the edges and run up against current limitations. But improvements will be made over time. More diverse representation of people, increasing the range of stylistic capabilities, and improving the ability to manipulate more original individual elements in imagery are necessary elements to move beyond experimentation toward stability and consistency.
So given the current constraints, how are creatives leveraging these capabilities today? One example is how visual artists are using tools like Midjourney to explore and iterate conceptual designs. They push their idea as far as they choose to take it with the assistance of the AI, and then output the final result as a high-res image asset. They take that image and use it as the foundation for a digital canvas, where they digitally paint over the scene, adding elements, removing or changing aspects, and refining the composition to achieve a final piece as an AI and human-codeveloped asset. These artists are building a bridge between current capabilities, traditional techniques, and the future of this evolving space. Think about the possibilities across other creative endeavors. Writing linear stories, connecting scenes while editing video, and even composing pieces of music might someday become faster and easier with an AI assisting to bridge between ideas and inspiration to finalized work. Encourage your teams to experiment.
We are in the very early days of what will likely be a revolution in how we leverage computing. So far, AI has been somewhat hidden and abstracted away from professionals and consumers alike. But now we are starting to see the emergence of AI and ML capabilities in the open, directly visible, and fully accessible to end users. Creative leaders acting as guides for their teams towards the future that AI will unlock is required for ongoing success. Helping individuals learn, adapt, and grow – at a continuously accelerated pace – is key for the advancement of your teams. It is also essential to the discovery of solutions to problems and opportunities we can not yet even anticipate.
Further Reading
The image entitled “A Bicycle for the Mind” was created with Midjourney using the prompt:
“imagine/ futuristic person riding electric bicycle, ralph mcquarrie style --ar 16:9 --test --creative --upbeta”