I recently listened to a fascinating talk by an AI expert(Standford' Professor Jeremy Utley) that completely changed the way I think about working with AI. This expert focuses on helping non-technical professionals collaborate effectively with generative AI. But the approach he shared was radically different from how we’ve traditionally thought about AI.
The Paradox of Asking AI How to Use AI
One of the most striking parts was this idea: “If you don’t know how to use AI, just ask AI.”
Unlike Excel, which can't teach you how to use Excel, or PowerPoint, which can't explain itself, AI is uniquely capable of teaching you how to use it.
The expert suggested a prompt like this:
“Hi, you're an AI expert. I’d like your help figuring out how I can best use AI in my work. As an expert, please ask me questions—one at a time—until you understand the full context of my workflow, responsibilities, KPIs, and goals. Then, give me both obvious and unexpected recommendations on how I can leverage AI.”
Apparently, this kind of conversation can be incredibly eye-opening—because AI is capable of self-reflection about its own capabilities.
The Inspiring Case of the National Park Service
There was also an amazing real-world example. The National Park Service had brought in this expert to teach 60 backcountry rangers and facility managers the basics of AI collaboration. Among them was a man named Adam Minor from Glen Canyon National Park.
He followed the advice, “Focus on the part of your job you dislike the most,” and immediately thought of the paperwork required every time he had to replace carpet tiles—paperwork that typically took two or three days.
He asked himself: “Could AI help me with this paperwork?”
In just 45 minutes, he built a natural-language tool that generated the required documents, saving two days per task.
But it didn’t stop there. Someone else accessed the tool and shared it across other parks. The U.S. National Park Service, which manages around 430 parks, now estimates that this simple tool—created in under an hour—could save them 7,000 workdays this year.
A non-technical employee, with just basic training, created something with massive impact.
The “Capability Gap”: Why Most People Don’t Benefit from AI
Studies show that AI can help people work 25% faster, produce 12% more output, and improve quality by 40%. Yet fewer than 10% of workers experience meaningful productivity gains from AI. This expert calls it the “capability gap.”
More surprisingly, studies in Europe and the U.S. found that AI doesn’t necessarily make people more creative. In fact, it sometimes makes them less creative. Upon deeper analysis, a key difference emerged between high-performing and low-performing users.
Tool vs. Teammate: A Difference That Changes Everything
Low-performing users saw AI as a tool, while high-performing users treated it as a teammate.
This mindset shift changed everything.
For example, when AI gives a mediocre result, how do you respond? If you treat AI as a tool, you might just discard or lightly edit the output. But if it's a teammate, you'd give it feedback, coach it, mentor it—just as you would a colleague whose draft wasn't quite right.
Those who treat AI as a teammate engage it in dialogue, asking it to ask them better questions, to clarify goals, or to suggest what they should be asking it.
Instead of: “Here’s my question, give me an answer,”
they say: “What are 10 questions I should be asking?” or “What do you need to know from me to produce the best answer?”
Practical Use Case: Roleplaying Difficult Conversations
One powerful example was using AI to roleplay difficult conversations with colleagues. You can feed it background on the person, have the AI construct a psychological profile, and then let the AI play that person in a roleplay. You can even receive feedback on your approach from that simulated perspective.
The expert called these “drills”—structured explorations that expand the boundaries of how we think AI can be used. He noted that his students often come up with use cases he never imagined, even after two years of teaching AI collaboration.
A New Understanding of Creativity
The speaker admitted he never considered himself a creative person, but working at Stanford’s d.school helped him see that every person is inherently creative—they just need help unlocking that creativity.
He shared a story about co-teaching with Grammy-winning hip-hop artist Lecrae.
They asked MBA students to “go out and find inspiration,” and when the students struggled, the expert asked Lecrae how he approached inspiration. Lecrae answered in one sentence:
“Inspiration is discipline.”
That moment revealed a key insight:
Many people don't even recognize inspiration as a tool they can cultivate. But the most creative people deliberately train their inputs because they know those inputs shape their outputs.
The same applies to AI:
“What kind of inspiration are you giving the model?”
Everyone has access to the same ChatGPT. The difference lies in what you feed it—your ideas, experiences, and perspectives. That’s what makes the output unique.
The Best Definition of Creativity
A 7th-grade student in Ohio once wrote on a Post-it:
“Creativity is doing more than what you first thought.”
The expert loves this definition because it cuts to the heart of cognitive bias—specifically functional fixedness and satisficing (settling for “good enough”).
True creativity is about pushing beyond that first idea.
Creativity in the Age of AI
When asked whether AI changes the definition of creativity, he answered: No.
Reaching “good enough” is easier than ever thanks to AI. But exceptional results still require iteration, input, and effort—just like always.
What’s changed is that our ability—or inability—to be creative depends more than ever on how we collaborate with AI, and what goals we pursue (explicit or implicit).
A Message to Creators
He emphasized that creators shouldn’t fear AI—they should embrace it. AI won’t replace them; it will liberate them in ways they’ve never experienced before.
His most powerful closing line?
“There is only one wrong answer to the question of how to use AI: ‘I don’t use it.’
I don’t use AI. I work with AI.”
When you begin to work with AI, everything changes.
The difference between using a tool and collaborating with a teammate is profound—and how we choose to view and engage with AI will completely reshape what’s possible.