The 12 Laws of AI – Law 11: The Law of Creative Exploration
January 17, 2024
The following is part of our blog series “The 12 Laws of AI.” The series is a set of practical and philosophical guidelines for DMOs to work from as they explore the opportunities and challenges of artificial intelligence (AI). Check out the previous blog in this series, “The 12 Laws of AI – Law 10: The Law of Democratized Innovation“
Law 11: The Law of Creative Exploration
We humans are both creators and explorers, delving deep into uncharted territories for moments of pure revelation: the Creative Spark. This profound clarity, born from myriad inspirations, is the hallmark of genuine innovation. True creative mastery in this new era will pivot from the creative skills of old to a deeper understanding of aesthetics, nuanced prompts, and an insatiable appetite for discovery.
AI couldn’t have created the blues
We, humans, are endowed with an incredible, extraordinary gift that really no other species on the planet is capable of — creativity. Creativity in the sense that we can create or bring things from nothing into the world: ideas, scientific advancement, new social constructs, Cheetos — you name it. And this is unique to us as a species.
Much of the fear about AI centers on the idea that it could replace humans as the dominant species one day. Science and technological advancements in the past 100 years have taught us that seemingly anything is possible. Still, at least as it currently exists, AI cannot replace us because, in its current form, it is utterly incapable of creativity in the true sense of the word. It cannot be created from pure nothing. ChatGPT can spit out 100 taglines for you if you ask, but those taglines will come from a combination of the input you provided, the constraints you’ve given, and the data it has been trained on. It’s not creating anything genuinely new so much as iterating on what we humans have already created before.
AI could not create the blues but could write a (likely bad) song. What do I mean by that? As a musical genre, blues music has its roots in the many musical subgenres created by enslaved African Americans during the mid-19th century. The blues was aptly named because it was an expression in the musical form of the daily hardships and tribulations forced upon this group of people.
The gentleman pictured above is blues legend and pioneer Robert Johnson. Johnson was known as the first rock star and helped to bring the blues into the mainstream during the early 20th century, despite a career that spanned just a brief seven months. If you have watched the Cohen brothers’ movie “Oh Brother, Where Art Thou?” you might remember Tommy Johnson, the Mississippi bluesman that the three main characters encounter at a crossroads in the Mississippi Delta. When the three men asked him what he was doing there, he said that he had been there to meet the devil. And they asked if he had. He said yes and that the devil had taught him to play the guitar in exchange for his soul. This story is derived from the legend of the real-life Robert Johnson!
Robert Johnson is a beautiful example of why humans have the ultimate trump card over artificial intelligence. Johnson’s life is a near-total mystery, but suffice it to say that his ability to create music was a mixture of immense talent and human experience. Conversely, AI can’t experience life, as Robert Johnson and his contemporaries did. It doesn’t know or understand heartbreak, hardship, joy, or other emotions, much less use those experiences as a foundation of creative expression.
AI platforms like ChatGPT can, however, create a blues song from scratch or even mimic the style of the outstanding Muddy Waters…but the resultant song is unlikely to be much more than a party trick rather than a chart-topper. It will simply be iterating on what we, as humans, have already created and cannot infuse its emotions and experiences, as described above, into a proper art form.
Despite its inability to truly create independently, AI can be a fantastic tool to help humans be creative. Remember the old scientific method we all learned in high school? It provides a robust framework for employing AI effectively in the creative process. If you recall your high school science class, there are six steps in the scientific method.
- Ask a question, then collect observations and background research.
- Form a hypothesis
- Test your hypothesis via an experiment
- Collect your data/results
- Analyze and draw conclusions
- Report your findings
We can use a slightly modified version of this framework to help us with many creative tasks. Let’s work through an example together.
In this instance, we will pretend the team at Travel Portland has asked us to help create tagline ideas that the organization will use to promote travel to their city to potential visitors residing in Dallas, Texas. Let’s explore how we could use the scientific method, beginning with Step 1, asking a question—or, more specifically, in our case, defining our task’s activity and objective.
Step 1: Define the task and our objective and ask a question
Our task is to create seed ideas for a new tagline to create a new one that effectively helps promote travel to Portland from Dallas. Our question is, “What tagline style will resonate with Dallas residents and encourage them to consider visiting Portland?”
This part of the creative process, where the whiteboard is a big empty space, and we’re not quite sure where to start, is where AI tools shine. Typically, in this part of the process, we would spend a chunk of time researching and learning a few essential items:
- What are Portland’s key strengths and appeal as a destination?
- What are our target audience’s predominant lifestyles and interests? Who calls Dallas home?
- Are there any current trends in travel or consumer spending or potential barriers to consider?
In addition to the usual research resources that an agency or DMO team would utilize to answer these questions, Google’s Bard can be a fantastic resource, thanks to its real-time connection to the web. For example, we could ask Google Bard to identify the top 10 destinations in the United States for visitor satisfaction; once we have those results, we can follow up by asking Bard which attributes Portland shares with those destinations. This is just one of many avenues to consider when using Bard as a research tool. Why not ChatGPT? Remember that ChatGPT is limited in its knowledge of the world, at least as of this writing, to mid-2021, so that platform won’t have the up-to-date knowledge we need for this particular task.
Step 2: Form a Hypothesis
Once we’ve defined our task and objectives, gathered our necessary research, and reviewed them, we can hypothesize what a strong tagline might need to incorporate. Our research informed us that Dallas residents love outdoor activities, beer, beautiful scenery, and quick culture — all things that Portland has in abundance! Taking that research and combining it with other insights and thoughts about what might help us create some good tagline ideas, we come up with the following hypothesis: a series of taglines that playfully contrast the differences between Portland and Dallas while highlighting why someone from Texas should visit will help us attract Dallas visitors to our city.
Now, we’re all set to begin collaboratively creating tagline ideas using AI tools like ChatGPT and Google Bard!
Step 3: Experiment/Ideate
This step is where the rubber hits the road and often represents the most time-consuming and intimidating step of the creative process: filling that empty whiteboard with seed ideas that can lead us to a final solution. Thankfully, now we can use our AI tools to get off to a running start. We have a pretty straightforward question — what would some effective taglines look like that would draw the attention of Dallas residents and interest them in learning more about visiting Portland? So now we need an effective command prompt to help us get the sort of output we’re looking for. Remember the framework we discussed in Law 8? Let’s use that here! If you’ll recall, there are five elements in putting together an effective command prompt:
- Include constraints
- Provide a role or framework
- Define the task
- Provide context and background
- Identify the steps, if needed
Here is a prompt we could use to generate our taglines to be used as idea seeds:
“Acting as a destination marketing professional ideating on a new tagline for a campaign promoting travel to Portland, OR from Dallas, TX. You will randomly generate ten words, five associated with Portland, Oregon, and five associated with activities and interests that might appeal to residents of Dallas, Texas. Next, you will generate a bullet point list of ten taglines by combining or associating 2 of the words you identified (1 about Portland and one about Dallas) and then using them as a humorous contrast to one another. Please consider that this will be a tagline promoting summer travel, including climate differences between the two cities.”
Step 4: Collect our Data/Results
The prompt above gave us the following results in about 20 seconds:
- “Melted in Dallas? Chill out with Portland’s cool vibes!”
- “Swap Texas BBQs for Portland brews because beer’s better than sunburn!”
- “Portland: Where summer doesn’t feel like you’re walking on the sun!”
- “Ditch the Dallas sizzle for a drizzle in the City of Roses!”
- “Portland: Where the beer is better than the BBQ. Just sayin.”
- “Texas turning you into toast? Jam out in cool Portland!”
- “Leave the Texas furnace, embrace Portland’s outdoor terrace!”
- “Portland: Where the only thing hotter than Dallas is our coffee scene!”
We only generated ten ideas for our example here, but if we’d asked for 50, 100, or even 500 of these, ChatGPT and Google Bard would happily comply. Try getting an intern to do that! These taglines are not ready for primetime on their own, but they supply the creative boost we need to create final examples that could be used in the Dallas campaign.
Step 5: Analyze and Draw Conclusions
Now that we have a few examples, we can analyze the potential taglines by utilizing ChatGPT’s ability to mimic focus group testing! Let’s write a command prompt to gauge the reactions to our favorite tagline.
“Acting as a DMO marketing research assistant, please create a table that displays first impressions of the tagline shared below, which will be used to draw interest from residents in Dallas about visiting Portland, Oregon. Please provide impressions from the following three personas: a 30-year-old single male professional from Dallas, a 40-year-old female professional married with two kids, and Yoda from Star Wars. The tagline is: Visit Portland, where the beer is better than the BBQ. “
Here is what the prompt returns back:
Pretty neat, right?! I think we might even have a shot with Yoda.
Here are just a few other examples of the types of analysis generative AI tools like ChatGPT and Google Bard can perform (ChatGPT generated this list; I simply asked what kinds of analysis of our tagline can be performed!):
- Semantic Analysis: I can break down the meaning and implications of each word in your tagline to ensure it conveys the message you’re aiming for.
- Sentiment Analysis: While I don’t have real-time sentiment analysis tools, I can give insights into the potential emotional tones conveyed by a tagline.
- Clarity Check: I can review a tagline to ensure it’s clear, concise, and free from linguistic ambiguity.
- Keyword Insights: Based on the language and phrasing, I can suggest relevant keywords or concepts that the tagline might evoke, which could be useful for accompanying content or SEO.
- Rhyme & Rhythm Checks: For catchy taglines, rhyme and rhythm can be vital. I can assess and suggest improvements based on these linguistic qualities.
- Preliminary Trademark Search: While not exhaustive or a replacement for a legal search, I can provide initial insights into whether a tagline is closely related to known trademarks or popular phrases up to my last update.
Step 6: Report Results
Utilizing the scientific method in our creative process, we now have a batch of solid tagline ideas to present to the Travel Portland team for further refinement and discussion. This same process could be utilized for any number of creative tasks like designing a new logo, selecting new branding colors, and creating website content, to name a few.
Other blogs in this series:
Law 1: AI is the Tool, Not the Craftsman
Law 2: Humble Beginnings and Law 3: Transparency
Law 4: The Law of the Imperfect Mirror
Law 5: The Law of Liberated Potential and Law 6: The Law of Collective Empowerment
Law 7: The Law of the Artful Inquiry
Law 8: The Law of Constructive Command
Law 9: The Law of Data Enlightenment
Law 10: The Law of Democratized Innovation
Law 12: The Law of Ascendency
Note: This collection of “laws” on AI incorporate insights from my research and writing on the topic. To make it as memorable as I could — and to demonstrate one of the many powerful utilities these tools offer — I asked ChatGPT 4.0 to style my writing in the voice of Robert Greene, author of the best-selling book “The 48 Laws of Power.” I hope you will agree that each of the Laws is a bit more memorable with this distinct style being employed. It’s crucial that we embrace these new tools and transparently acknowledge how they improve our critical thinking and public sharing of ideas.