AI Needs Strong Human Talent

Unleashing your creativity has never been easier with the advent of design and copy AI engines like MidJourney, Leonardo, and Chat GPT. These incredible tools have the potential to supercharge human creative output, making them indispensable assets for businesses.

However, the idea of AI replacing human talent en masse anytime soon seems far-fetched. These engines require the input and direction of skilled professionals to maximize their potential. Rather than replacing human talent, they will eventually enhance the capabilities of existing teams, allowing them to uplevel their creative capabilities, increase productivity, and deliver superior work. Alternatively, ignoring them may not hurt you immediately, but will eventually leave your organization behind.

In this article, I'll share six observations on where AI needs human assistance and how these human/machine partnerships can be improved. But, I’m new to this world and am learning. So, share your point of view, and for those confident that AI design tools are game-ready, there’s a challenge for your below.

1) Garbage In, Garbage Out: Dust off your creative brief writing skills because AI will depend on them even more than the creative team that generously helps turn so-so briefs into strong marketing campaigns. AI engines may not be so forgiving. These engines deliver the copy and visual elements you ask for incredibly fast, but they depend, quite literally, on your job description (aka the creative brief) to generate their work.

Requests that lack clarity or detail deliver weak or off-target results requiring more cycles to improve. Overly specific requests are equally challenging since they require much more detail and an alignment between what the requestor wants and the language the engine needs to deliver that job. Because of this, getting what you ask for from an engine will require the requestor to be in tune with the AI’s underlying language. Like any language, speaking AI-ese fluently will require expertise, and creative teams seem to be the most natural choice to lead that effort since they already translate current business needs into creative solutions for the organization.

Since these tools will develop rapidly, don’t wait. Encourage your creative team to start experimenting with AI tools now so they can learn how to translate your needs into AI requests that deliver better results.

2) The Concept is Critical: Strong and lasting marketing campaigns are concept-driven. The concept helps make the work unique, attention-grabbing, and repeatable. But engines will take your request at face value and deliver it. Unless your well-written job request also provides the A+ concept you will end up with nice-looking, forgettable, limited work.

The need for great concepts will become truer as more companies adopt these engines. As they do, visual and copy layers will generally improve, making a strong concept increasingly important to stand out and break away from the prettier ephemera.

In my experience, the creative team’s ability to transcend a brief’s literal meaning and find more interesting ways to express the root idea has always been at the heart of its value. As long as these engines remain rooted in literalness or only seek to build a single job without an eye on the broader campaign they are insufficient. Make sure your organization has strong senior creatives that lead more complex and valuable projects.

3) The Artist’s Imagination: During the pandemic, I wrote a children's story for my son's school, but I knew that the tale needed illustrations to bring it to life. Fortunately, I discovered Deborah, a fine artist on Upwork who did an incredible job. Despite only giving her the book copy she seemed to envision the story with ease immediately, requiring only a few clarifications to nail the illustration style I was looking for. What’s more, along the way she added unique elements based on her own research that went beyond my brief and improved each illustration.

As an experiment, I attempted to use Mid Journey, an AI design tool, to generate an image from the book. However, the challenge of creating a single character from the textual descriptions proved difficult enough, let alone crafting 15 full scenes that worked together and complemented the story. While my resulting character was certainly a vast improvement over my own innate artistic skills, it lacked the intricacies and nuances that brought Deborah's illustrations to life. Instead, my character looked like a flat figure superimposed on a rich background.

As I reflected on my experience, it became clear to me that while AI tools are improving, they have yet to replicate the imaginative power of a skilled artist. Unless one is proficient in using these tools, it's best to rely on the talents of artists who can see beyond the literal request and deliver more interesting work

CHALLENGE: I am just starting my journey with these tools and would love to be educated. If you can build better illustrations than what I received from Deborah, I’ll pay you the same rate that I paid her for an illustration. If you’re interested in taking the challenge, contact me at Jim@bobolonpress.com and I’ll share the details.

4) Quantity or Quality?: Check out Alice and Sparkle, a book created by Ammaar Reshi using 100% copy and design AI. The main character is adorable and the quality of this book is quite good, considering it was done on a weekend by a machine.

I’m envious of the time savings AI offered Ammaar. But the work also looks familiar, is lacks some consistency from page to page, and has some weird flaws. For example, that adorable Pixar-style little girl has claws on some pages.

If your work requires generating lots of decent work as quickly as possible, full steam ahead. But in a world with billions of people supersaturating it with new ideas, how do you expect to get noticed by being average? We don’t flock to bands with average talent. Professional sports aren’t filled with average skilled players. Likewise, our creative standards should not be diminished just because we can get more stuff faster.

Does this mean these tools are useless? Heck no. They can be profoundly helpful during ideation and the early phases of the creative process. They can allow your team to choose from more options and see them with greater clarity. But getting the most out of them will require human judgment and scrutiny to push work further or reduce low-quality work so that what emerges is sufficiently strong to represent your brand.

5) Humans like Consistency. Although we want our work to stand out, the fact remains that consistency of message, style, format, and visuals aids the customers’ understanding of your message and leads to higher brand awareness. If engines approach each job as new and fail to build consistency, you run the risk of your work becoming too diverse, not connecting, and limiting customer comprehension.

To truly leverage the power of AI engines, I think companies will eventually maintain databases of creative and marketing work overlayed with essential data that hones the reference material your AI uses. By doing this, these engines can learn from previous work and evolve toward what’s working best, even as it adapts to changing circumstances. It can also be used to help keep work more consistent by locking in brand elements.

To my knowledge, these databases don’t exist yet, but eventually, they will need to be built and managed so the incredible power of AI can be focused on your company and the brand you are trying to define. Although many people may influence this work marketing and creative staff are great candidates given their connection to the original work and data. Regardless, think of your previous creative campaigns as data that AI engines can use to hone their work for your organization and bias its output based on what worked previously.

6) Know thy customer. AI-driven engines extract from the world, everyone. But unless you’re ubiquitously appealing or have unlimited marketing funds, you market to a much narrower audience. These narrower audiences are usually identified through research and once discovered help guide product and marketing decisions. This targeting is essential to getting the most return from your marketing investment.

So what we need is, like above, AI that can be steered to look more narrowly. To do this I imagine companies will build persona databases, like the creative database, that influence your AI’s sources and lead to more tailored output. These databases will allow AI engines to look more specifically at your customers for language and trends that will resonate with them, identify cultural events in real-time that are affecting them, and use this information to build work that better connects with them. If maintained, they can supercharge relevancy at scale and at superhuman speeds.

Once again, however, they will need to be built and maintained and marketing and creative teams seem appropriate since they already think in terms of target audiences. Getting to this level will take time, but eventually, users will need to focus the power of these engines on the audience we care about most.

So, will we continue to see more output from these AI engines? Of course. They are shockingly good already and will only improve with time. They will change the way we interact with the world and soon be embedded in more of our everyday interactions.

Is that a good thing? Probably. They provide increased capabilities at superhuman speeds, but they are not perfect. For the foreseeable future, they require human partnership and expertise to use them best. But change is coming, and rather than wait for that change to arrive, begin your journey with these capabilities now. Challenge your creative teams to learn and use them so they can eventually deliver even more high-quality work faster.

My thoughts on these topics are developing, so I encourage your thoughts and comments, send me a note: jim@bobolonpress.com.

Jim Bobowski

Jim is a marketing consultant that specializes in marketing strategy, product positioning and driving growth at startups and established companies.