The 12 Laws of AI – Law 8: The Law of Constructive Command
November 29, 2023
The following is the sixth part in 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 7: The Law of the Artful Inquiry“
Welcome back to the 12 Laws of AI! In Law 8, the Law of Constructive Command, we’ll learn about a framework that makes working with AI tools and getting the responses desired much more manageable.
Law 8: The Law of Constructive Command
To navigate this new frontier, your ability to command AI with knowledgeable precision is paramount. Every detail, from the very words to the subtle nuances, forms the parameters of your expedition. To navigate with precision, command with deliberate intent. If the path grows dim, adjust your bearings. Should you crave more profound revelations, invoke the wisdom of noted guides or established frameworks. Leaders will realize the art lies not just in the asking but in the artful crafting of the command.
“Garbage in, garbage out.”
George Fuechesel, IBM
In Law 7, we learned how to ask better questions — an essential skill to improve your ability to take advantage of the many forms of AI available to us. However, the biggest hurdle remains — the empty prompt staring at us whenever we open our favorite chatbot app or AI tool. As your early explorations with ChatGPT or Google Bard likely showed you, it can still be a challenge to know how best to structure your query or question to ensure a response that fits what you had in mind. Are you looking for a concise 300-word summary of an article you just read? Or would it be better in a bullet format? Do you want these notes to consider that you are a DMO professional and expand the summary to explain how the material might relate to your organization? Well, with apologies if you’re not a Portlandia fan…we can pickle that (if you’re not familiar with the show or that particularly hilarious episode, it’s worth a watch on YouTube)!
Before we discuss the framework that will make your artful inquiries come to life, remember: AI tools, especially chatbots, have been created to be interacted with in natural language, i.e., the same language you use to interact with other humans. It can even identify slang and understand your typos without breaking much of a sweat. So, don’t worry too much about nailing perfect grammar or formal tones.
Our framework relies on five easy steps:
- Include Constraints
- Provide a Role and/or a Framework
- Define the Task
- Provide Context and Background
- Identify the Steps, If Needed
Let’s discuss each step in turn:
Include Constraints
Especially with chatbot AI tools like ChatGPT and image generation tools that allow for enormous flexibility in shaping responses to queries, it’s critical to inform the AI precisely what format you’re looking for, the desired length, and even the required tone or complexity.
For example, if we asked ChatGPT to summarize how a DMO would incorporate the primary ideas from Seth Godin’s seminal book “This is Marketing” into their marketing plan, by default, you’ll likely get an in-depth, multi-paragraph response that will stretch somewhere between 500 to 1000 words in length. Beefy! But if all you wanted was a simple list of bullet points or a summary that would be easy to understand for anyone to digest, you’ll want to add appropriate lanes for the AI to stay inside of as it crafts a response. So, for example:
“Using an informal tone, please provide a 100-word, one-paragraph summary of the book “This is Marketing” by Seth Godin and explain how the book’s primary ideas could be utilized in a DMO strategic marketing plan.”
The resulting response will provide a summary to hit the high points without overloading you with the details.
“Incorporating Seth Godin’s “This Is Marketing” principles into a DMO’s strategy involves focusing on exceptional aspects of the destination, targeting niche audiences, using permission-based marketing, storytelling, solving real problems, and offering value. Consistency, empathy, and connection are key to building trust over time. This approach creates more meaningful, effective campaigns, fostering visitor engagement and loyalty.”
Would you prefer a table summary? No sweat. Just adjust the command to provide the summary as a table, and ChatGPT, in this example, will take care of the rest:
Exploring how to incorporate critical ideas in books you’ve read into your organization’s plans provides a near-limitless well of new ideas to consider. We’ll discuss this potential for exploration in a later law, but constraints will ensure that your AI sidekick only gives you as much, or as little, as you are looking for and in just the right format for your learning style or purposes.
Provide a Role and/or a Framework
ChatGPT and Google Bard make for incredible assistants. Both can assume roles like a DMO professional from California (yes, that specific if you need it!) or a marketing layperson (helpful in ensuring that query responses are more digestible for general audiences).
The best way to take advantage of this capability is to include “acting as if” in your query, for instance:
“Acting as if you’re a California DMO CEO, and in bullet point format, please explain how recent economic conditions might affect travel to smaller inland destinations in California. Please include citations as you can.”
The resulting response will give you a terrific breakdown of the primary economic factors affecting travel to our example destination and even include links to pertinent information sources supporting each bullet point. Note: ChatGPT can provide a good response here but is limited to events on or before 2021, so Google Bard is the better choice of platform in this instance.
Finally, you can define a framework for the AI to follow to ensure you get the desired response to a specific ask. Our example above, where we asked for a strategic marketing plan that adapts the key ideas of Seth Godin’s book “This is Marketing,” is a prime example of putting this idea to work. Another example would be to include details of your destination and then ask to create an outline of a plan incorporating the ideas in Godin’s book — or even a collection of books or articles you’ve read. You can include as much in your framework as needed.
Define the Task
It might seem obvious, but including what you want the AI to do is essential. In our first example, recall that we asked our AI assistant to summarize how a DMO could use the fundamental principles from a specific book in a strategic marketing plan. That defines the task and ensures we get a proper response. Remember that you can be as detailed as you need, the same as when determining the role or framework for the AI to work within. Do you need specific tactics and an overarching strategy for a summer marketing plan that stays true to Godin’s book? Just update the task you’re asking the AI to complete, and you’re off to the races. Remember to consider AI responses as terrific starting points for developing a fully fleshed-out strategy. AI tools like ChatGPT haven’t yet replaced humans in creating nuanced and complex plans to ensure success in a crowded marketplace. Still, this is tough to beat for brainstorming and getting off to a fast start in the planning process!
Provide Context and Background
ChatGPT and Google Bard’s awareness of our world and how it all ‘fits’ together is remarkable. Granted, ChatGPT is limited to a world view from 2021 and earlier, but it still can show fantastic capacity to adjust a response, given the context and background you provide in your query. Including details about your destination, a desire to take into account current trends, or the background of the reason for your query can result in much more textured and nuanced responses. You can include a request for the AI to consider current travel trends and economic conditions (For example: “Given that the economy is slowing and sustainable travel is becoming more popular, how should we adjust the strategy outline above?”), what are your specific goals for the requested strategy (For example, to increase brand awareness), and much more?
Identify the Steps (if Needed)
If you’d like the AI to follow a specific sequence in providing a response or structure the output a certain way, just add those steps at the beginning or end of your query. For instance:
“…you’ll summarize the plan in no more than 500 words, arrange it in up to 3 paragraphs, then outline the plan’s steps in bullet point format.”
“You are to act 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 that might appeal to residents of Dallas, Texas. You will generate 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.”
That last example is a doozy, right? But fear not; the AI will have no problem breaking down the elements of your request and will even automatically provide new prompts asking for your following inputs as you have directed. The ability to give a series of inputs and related outputs is a powerful way to supercharge how you collaborate with AI tools, enabling you to take on more complex tasks or develop more profound brainstorming ideas in a far more efficient manner.
There you have it! Five simple steps you can use to significantly improve your AI experience and ensure that you’re getting the responses you want and enabling some abilities you might not have known the AI could accomplish. We’ll put this technique to work in the laws to come, as we get into some fascinating specific use cases, including data analysis, the creative process, and even product development.
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 9: The Law of Data Enlightenment
Law 10: The Law of Democratized Innovation
Law 11: The Law of Creative Exploration
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.
