Keeping It Human in a Digital World (8)

I call the A3 Framework a learning journey for a reason. AI is so different from the software we’re used to, it’s often hard to grasp how it can help us at first. That’s why approaching this work as a student with a “beginner’s mind” is crucial. When we make this small mental shift, we tend to allow ourselves more space and time to cultivate curiosity.

Through my research, I’ve observed that during this exciting and sometimes unsettling time of rapid AI advancement, those who thrive with AI tend to:

• Make the leap to experiment with AI even before they can clearly articulate its value.
• Keep a curious and open mind when they jump into the strange new world of human-AI communication.
• Invest focused time in experimentation.

While much about AI remains unclear, engaging now sets you up for compounding benefits as the technology evolves. You’ll not only experience immediate gains in productivity and performance, but by developing a foundational understanding of AI, you’ll be better positioned to grasp future developments and understand their impact. Essentially, by prioritizing learning today, you’re building AI muscle that will accelerate your growth tomorrow.

Before we dive deeper into the A3 Framework (Chapters Thirteen through Fifteen), we will first explore some basic principles that will help you succeed in this strange new world of interacting with AI. We’ll begin by examining the fundamental shift AI introduces in our relationship with software (Chapter Eleven), followed by key principles of AI communication (Chapter Twelve).

11. It’s Time to Rethink Your Relationship with Software

Whether you’re just starting out or you’ve been in the field for a decade, interacting with today’s AI systems is unlike anything you’ve experienced before. The technology we now have at our fingertips advanced so rapidly that it surprised even its creators. AI pioneers dreamed of this day, but few imagined it would arrive so soon or be so widely accessible to digital citizens the world over.[i] We are now living in a moment in which individuals have access to nearly the same quality AI as corporations do, but all of us—individuals, teams, and corporations—need to learn how to use it well.

AI Redefines What Software Can Be

To get value from AI, we first need to understand that this is a whole new world of software that requires us to interact in fundamentally different ways. Our existing mental models for working with technology don’t help us understand how to engage with AI or define its role in our work. Traditional business processes, built around conventional software, simply don’t map to this new paradigm of collaborating with machines. Interacting with AI can be a somewhat messy and strangely human-like process, but this is also what makes it feel almost magical.

In fact, what makes AI “fail” as traditional software is exactly what gives it superpowers that allow us to use software in ways we never could before. This is the AI paradox: while unreliable at the precise tasks we traditionally turn to software to perform, the machines’ human-like “imagination” (combined with its superhuman processing capability) is what gives us powerful new tools for creative automation.

Predictably Unpredictable AI

Conversations with AI can feel, at times, eerily human-like. These models not only return responses that historically could only be composed by a human; they can also be quirky and inconsistent. While we are used to our friends, family members, and colleagues having “off” days, we certainly don’t expect that from our software. Sometimes, in a way that is not dissimilar to a human teenager (at least the ones I have in my house), your AI can seem to have different moods and even personalities from one day to another. Like a human, when we coach it—or strangely, if we even cajole or negotiate with it—we get better outcomes. This is far from our experiences with accounting or customer relationship management software!

Traditional software executes precisely and predictably to predefined instructions. It has static, fixed capabilities and, if well-built, operates consistently, delivering identical responses every time we use the same query.

AI, on the other hand, is adaptive and evolving. It gives different responses each time, learns your needs, and works best when you dialogue with it.

This inconsistency is one of the reasons we often struggle to fit AI into our established ways of working. AI can have emergent behaviors that even its creators find difficult to predict, making it challenging to integrate into any process where we need consistent outcomes. The systems may even “lie” to you—phenomena known as hallucinations. This happens because AI is designed with the goal of providing responses that are helpful and relevant to your requests. In trying to meet this goal, it takes its interpretation of your request to guide the search through its vast training data—and then uses mathematical probabilities and patterns to find its own path to an answer. In the process of trying to find a “best” answer, it may make something up and confidently and convincingly present these fabrications to you as fact.

Defining the Undefinable

It’s no surprise that everyone seems to be working to wrap their head around what this new thing is, what role it should play, and how to use it. AI isn’t like traditional software, and it’s certainly not human. So, what is it?