Malpractice

I'm standing in the Nikkei Business Press offices in Tokyo. Hair and makeup done. TV cameras positioned. I've been coordinating with Tanaka-san, the director, for weeks - back and forth on plans for a special edition AI bootcamp we're launching for the Japanese market. They've got their vast distribution list of subscribers ready to go.

I'm thinking we're here to workshop some ideas, maybe sketch out some promotional concepts on a whiteboard.

Tanaka-san looks at me and says: "Okay, Jeremy. Give me your killer line."

I freeze for a second.

Here's the thing - and maybe this is confession time - I felt embarrassed by what I did next. Because the prompt "give me your killer line" carries an assumption: that I have one. That I know exactly what to say. Implication being, if I were a real professional, I could just deliver on cue.

But the truth? We can hold very little of our knowledge in short-term memory, available for immediate delivery. Especially when it needs to be really, really good. Especially when it's going out to hundreds of thousands of people.

So I pulled out my phone.

"Tanaka-san," I said, "I'm going to have to invite my friend Claude to this conversation."

He looked confused. "What do you mean?"

"If you're asking for a killer line that you're about to blast out to hundreds of thousands of people," I said, "I wouldn’t entrust that responsibility to my gray matter. That would be malpractice."

I opened Claude - which I've trained on all my blog posts, podcasts, you name it - and said: "Hey Claude, I'm with the Nikkei Business Press team here in Tokyo. They want to record a one-minute PR segment. Based on what you know about my expertise and unique perspective on AI transformation, would you interview me like a PR director whose responsibility is to diligently prepare my cue cards? Please ask me a few questions so you can propose 20 hooks for this piece."

Claude asked. I answered. Twenty hooks appeared on my screen.

I showed Tanaka-san my phone. "Which angles do you like?"

He pointed. "Number 4, number 7, and number 14."

"Hey Claude," I said, "could you stitch 4, 7, and 14 together into a coherent hook? Max 300 words?"

Thirty seconds later, I sent it to the teleprompter. "Okay, Tanaka-san. Ready for my killer line?"

He stared at me for a moment. Then he said something I'll never forget:

"You know what's funny, Jeremy? We've been talking all this time about teaching people how to work with AI. It never occurred to me that we would work with AI."

That's the paradox we're living in. Everyone teaching it. Nobody doing it.

And here's what I realized in that moment of self-consciousness, when I felt like I was admitting I didn't have all the answers:

The word isn't "need." The word is responsibility.

It would have been malpractice for me, of all people, not to work with AI on something this important. But I still felt weird about it. And if I'm feeling that way, what about everyone else?

What Tobi Just Changed

Six months ago, Tobi Lutke (the guy who turned a snowboard shop into Shopify, a hundred-billion-dollar company) decided he was done with the paradox.

He sent a (now-viral) memo to the entire company. It leaked. And it's the clearest articulation I've seen of what's actually at stake here.

"Reflexive AI usage is now a baseline expectation at Shopify."

Not encouraged. Not recommended. Expected.

He continued: "Using AI effectively is now a fundamental expectation of everyone at Shopify. It's a tool of all trades today, and will only grow in importance. Frankly, I don't think it's feasible to opt out of learning the skill of applying AI in your craft."

Then he did something remarkable: He added AI usage questions to performance reviews. He mandated that before any team asks for more headcount, they must demonstrate why they cannot accomplish their goals using AI. He dedicated time to AI integration in monthly business reviews.

And then this line: "Everyone means everyone. This applies to all of us - including me and the executive team."

Tobi's not talking about adoption. He's talking about reflexivity. And there's a massive difference.

The Knowledge Worker’s Stethoscope

Here's a quick thought experiment.

You have a heart issue. You go to the cardiologist. You lie down on the table. The doctor walks in and puts her ear directly on your chest.

How would you feel?

You'd ask where her stethoscope was. Where's the MRI? Where are the diagnostic tools that let her do her job properly?

And if she wiggled her fingers, bragging, "Nope, only handmade, artisanal cardiology here!" you'd ask for a referral. Immediately.

The same is true of a radiologist who insisted on eyeballing your X-rays instead of using imaging software. We'd call that professional negligence. We'd call that malpractice.

So why do we accept it from knowledge workers?

If you're writing without AI feedback, analyzing without AI assistance, strategizing without AI as a thought partner, then you're the cardiologist with her ear on my chest.

Not bringing AI to your work isn't humility. It's hubris. It's the assumption that your unaugmented cognition is sufficient for professional-grade output in 2025.

It isn't.

What I've Learned From Thousands of Professionals

I've been surveying everyone who goes through my Personal AI Bootcamp - we're talking thousands of professionals at this point. And here's what they tell me about why they're not using AI more consistently:

It's not "I think it's dangerous." It's not "My boss won't let me." It's not "It's not useful."

The dominant themes are:

• "I simply forget to use it"

• "It's not part of my current workflow"

• "Old habits die hard"

This is a behavioral problem, not a knowledge problem.

And here's the vicious cycle I keep seeing: Low use leads to low comfort. Low comfort leads to low confidence in results. Low confidence leads to even lower use.

People don't know how, so they don't try. So they become people who "can't." So they stop trying altogether.

The only way to break this cycle is repetitions. You cannot tell someone to be comfortable with AI. You must force the reps that create comfort.

Which brings me back to Tobi's word: reflexive.

Reflexive doesn't mean you use AI sometimes. It doesn't mean you use AI when you remember. It means AI is your first move. It means the pathway from "I have a task" to "AI can help with this task" is automatic.

A good friend who runs a billion-dollar business unit for one of the largest privately held companies in the US has a post-it note on his screen: "Did you ask ChatGPT?"

That's a trigger. That's how you build reflexivity. Not through conviction, but through activation.

Baseball, or Grenade?

I was speaking at a conference at Adobe in Las Vegas recently. I held up the presentation clicker and said: "Imagine I'm going to toss this into the crowd. First person to catch it gets to come up here and share how they're working with AI in their role."

I paused.

"Here's my question: Are you thinking about this clicker like it’s a baseball? You're in the outfield, you've got your mitt ready, you're fighting for that pop fly?"

Another pause.

"Or is this like a grenade? And you're ducking, hoping it doesn't land near you?"

The room got quiet.

"Because for most of you," I said, "I think it's the grenade. And that's the problem. This isn't about your company. This isn't about your team. If you're not fighting to grab that remote like a bridesmaid at a bouquet toss, what are we doing?"

So let me ask you the same question. Honestly.

Where are you?

The Reflexivity Self-Assessment

Here's how you know if you're actually working reflexively with AI. Don't ask yourself "Do I use AI?" That's too vague. Ask yourself these specific questions:

Last email you sent: Did you use AI to draft, refine, or check tone?

Last presentation you gave: Did you get AI feedback on clarity, flow, or impact?

Last performance review you wrote: Did you leverage AI to ensure fairness, specificity, and actionability?

Last difficult conversation you had: Did you practice with AI first?

Last research project: Did you use AI to synthesize findings or identify patterns?

Last time you were stuck: Was AI your first call, or your last resort?

Last brainstorm: Did you use AI to generate alternatives before settling on your idea?

Last decision you made: Did you pressure-test your thinking with AI?

Last document you read: Did you use AI to summarize key points or extract action items?

Last creative brief you wrote: Did you ask AI to identify gaps or suggest improvements?

If you're answering "no" to most of these, you're not working with AI reflexively. You're working the way knowledge workers worked in 2019. (For more on this paradigm shift, check out Brice Challamel’s fabulous recent guest post, “Don’t Be Fred”)

And in 2025, that's malpractice.

The Only Right Answer

Tobi put it plainly in his memo: "Stagnation is slow-motion failure. If you're not climbing, you're sliding."

Here's how I'd say it: The only acceptable answer to the question "Did AI help with your work?" is "ABSOLUTELY."

Anything else is malpractice.

Not because AI is magic. Not because it solves everything. But because choosing not to use available tools that demonstrably improve your work is professional negligence.

You wouldn't trust a doctor who refused to use a stethoscope.

Why should anyone trust a knowledge worker who refuses to use AI?

The embarrassment I felt in Tokyo - that self-consciousness about pulling out my phone - was misplaced. The real embarrassment would have been delivering a subpar hook because I was too proud to ask for help.

Reflexivity isn't weakness. It's responsibility.

So here's my question for you: Can you screen-share right now and show me five ways you're working with AI? If not, you don't have the right to tell anyone else in your organization to do differently.

And if you can't answer "ABSOLUTELY" to whether AI helped with your last piece of work?

That's not a gap.

That's malpractice.

Want to build reflexive AI habits with your team? I regularly run enterprise AI bootcamps for organizations serious about making this shift. And if you're looking for tactical examples of how to integrate AI into your daily work as an individual, subscribe to my free AI newsletter or the aforementioned Personal AI Bootcamp.

Join over 27,147 creators & leaders who read Methods of the Masters each week