Why Human Creativity Is Your Best Defense Against AI in 2026 | Surviving AI

Think AI is going to take all the jobs? Think again. 83% of business leaders believe AI will elevate the importance of uniquely human skills — not replace them. And 60% of executives say human creativity remains essential to competitive advantage, even as AI handles the routine work.

But here is the paradox: as AI gets more capable, human creativity is actually declining. Students show a 42% decrease in divergent thinking. This episode makes the case that your creativity is not a relic of the pre-AI world — it is your ticket to thriving in the post-AI world.

In this episode, you'll learn:

• Why 83% of business leaders say AI will make human skills MORE important
• The science behind what AI genuinely cannot do: true creativity, empathy, and original thought
• The creativity paradox: why humans are becoming LESS creative as AI improves
• The economics of human creativity in an AI-dominated workplace
• Six actionable strategies to develop your irreplaceable creative edge
• How to position creativity as a career moat in the age of algorithms

Subscribe to Surviving AI and leave a review — it helps other workers find this show.

Surviving AI podcast, human creativity AI, future of work creativity, AI-proof skills, human skills vs AI, Carlo Thompson, creativity career advantage, AI can't replace creativity, divergent thinking decline, uniquely human skills, creative career strategy, AI workplace adaptation, soft skills AI era, career differentiation 2026

YouTube Episodes

SURVIVING AI With Carlo Thompson - YouTube

[00:14] Introduction: Addressing the "3:00 a.m. feeling" that AI is coming for creative and strategic roles.

[01:24] Unexpected Trajectory: Why AI is tackling art, coding, and strategy instead of just "dirty or dangerous" tasks.

[02:10] The Executive View: Analyzing data showing 83% of leaders believe human skills are actually more important now.

[03:11] Role vs. Task: The critical distinction between automating a specific activity and replacing an entire job description.

II. The Economics of the "Human Premium"
[04:18] The 30% Free-Up: How AI automates drudgery to free up time for high-value creative work.

[05:36] Super Agency: Moving from being a "laborer" to an "architect" of outcomes.

[06:18] Real-World Hiring Trends: Why non-technical, human-centric roles are seeing a massive spike in demand.

[07:31] The Jagged Frontier: Understanding where AI is a genius and where it fails like a 5-year-old.

III. The Creativity Paradox
[08:33] Humans vs. GPT-4: Discussing the study where AI crushed humans in standardized divergent thinking tests.

[10:44] Combinatorial vs. Transformative Creativity: Why AI is a "remix engine" but cannot achieve "World-Making" or the "Aha!" moment.

[11:56] The Epiphany Gap: Why AI cannot recognize when its own rules are wrong—the "Square Peg" problem.

IV. The EPOC Framework for Career Safety
[14:06] The Framework Defined: A guide to the five uniquely human pillars.

[14:30] E - Empathy: Reading subtext and high-stakes emotional undercurrents.

[15:08] P - Presence: The economic value of physical networking and trust.

[15:25] O - Opinion: Ethical judgment and being a "steward" of brand values.

[15:56] C - Creativity: Transformative leaps that don't exist in previous data.

[16:10] H - Hope: Visionary leadership that rallies teams toward the "impossible."

V. The Danger of Cognitive Atrophy
[18:38] The "Wall-E" Risk: How "cognitive offloading" is leading to a decline in human problem-solving skills.

[19:14] Homogenization of Ideas: Why relying on the same AI models leads everyone to "regress to

Transcript

SPEAKER_01 0:00

Artificial system on it.

SPEAKER_02 0:14

Welcome back to the deep dive. I'm your host, Carlo Thompson, and this is Surviving AI. Before we get started, do me a favor, hit that subscribe button, smash the like, leave a review.

SPEAKER_00 0:25

It really does help.

SPEAKER_02 0:26

It does. It helps the algorithm find us, which, uh, considering today's topic, is a little ironic. Because today we are tackling the absolute biggest elephant in the room.

SPEAKER_00 0:37

It's a pretty big elephant.

SPEAKER_02 0:38

It's massive. I'm talking about the anxiety that I think we are all feeling, whether we you know admit it at dinner parties or not. It's that gnawing 3 a.m. feeling that artificial intelligence isn't just coming for jobs in the abstract. It's coming for my job. It's coming for your job.

SPEAKER_00 0:57

Right.

SPEAKER_02 0:57

And for a long time the narrative was, oh, don't worry, you know, the robots will take the dangerous stuff, the dirty stuff, the repetitive stuff.

SPEAKER_00 1:03

Sure, they'll drive the trucks and flip the burgers. That was the idea.

SPEAKER_02 1:06

Exactly. But that's not what we're seeing, is it?

SPEAKER_00 1:09

No, the trajectory has been well, unexpected.

SPEAKER_02 1:13

Unexpected is a polite word for it. We are seeing it come for the writing, the strategizing, the coding, the art, the stuff we thought was safe. The human stuff. The stuff we thought made us human. And honestly, for this deep dive AI dispatch, creativity is your superpower. I came in ready to be depressed. I see an AI write a poem or design a logo in three seconds, and I start to sweat.

SPEAKER_00 1:35

I get it.

SPEAKER_02 1:36

So I need you to level with me. We've a massive stack of research today. Workday, MIT, McKinsey, LinkedIn. Is the doom and gleam warranted? Or is this creativity thing actually a lifeline?

SPEAKER_00 1:48

Well, if you look at the actual data, and you're right, we have a massive stack today. The story is actually quite different from the panic. In fact, if you talk to the people making the decisions, the executives and business leaders, they aren't looking to replace you.

SPEAKER_02 2:01

They aren't, because it certainly feels like they might be. If I'm a CEO and I see a tool that can do a task for, what, a fraction of a cent, why wouldn't I replace the human?

SPEAKER_00 2:10

So let's look at the numbers. There was a comprehensive study involving Workday and other major data sets revealing that 83% of business leaders believe AI will actually elevate the importance of uniquely human skills, not replace them.

SPEAKER_02 2:24

83%.

SPEAKER_00 2:25

And to add to that, 65% of executives explicitly started that human strategic decision making, intuition, and creativity are more essential now for competitive advantage than they were before the AI boom.

SPEAKER_02 2:38

Wait, let me pause you there. I need to untack that. 83% say human skills are going to be more important. That feels completely counterintuitive.

SPEAKER_00 2:46

It does.

SPEAKER_02 2:46

It feels like corporate speak. Like, if I have a machine that can do the work, why do I need the human? Are they just saying that to keep morale up before the layoffs start?

SPEAKER_00 2:55

It's a fair skepticism.

SPEAKER_02 2:57

Yeah.

SPEAKER_00 2:57

But the disconnect comes from how we, the public, view work versus how a business leader views work.

SPEAKER_01 3:03

Okay.

SPEAKER_00 3:03

We tend to view AI as a replacement for the role. I am a copywriter, the AI writes copy, therefore I am gone. The data suggests we should view it as a replacement for the task. The task. And those are very different things.

SPEAKER_02 3:15

Okay, let's dig into that distinction: task versus role. Because I think that is where a lot of us get tripped up.

SPEAKER_00 3:21

That is the mission of this deep dive. We are going to explore why this isn't an AI apocalypse, but a transformation. We're going to look at the hard scientific proof, and I mean rigorous peer-reviewed studies that show exactly where AI hits a hard ceiling in creativity. A ceiling. Good. We'll also talk about why using AI might actually be making us dumber if we aren't careful, the whole cognitive atrophy problem.

SPEAKER_01 3:45

Oh, great.

SPEAKER_00 3:46

And finally, we're going to give you a roadmap. We're going to break down something called the EPOCH framework and specific strategies to build what's called the cybernetic teammate.

SPEAKER_02 3:55

Cybernetic teammate. I love that term. It sounds very sci-fi, but I have a feeling you're going to tell me it's actually very practical.

SPEAKER_00 4:01

It is. It's about moving from doing the work to uh directing the work.

SPEAKER_02 4:07

All right. Let's dive into the first big chunk of our research stack. You mentioned this idea of transforming roles versus eliminating them. I've got the McKinsey research here in front of me, and they are pretty specific about this task versus job distinction.

SPEAKER_00 4:21

Right. So McKinsey's analysis, which is backed up by similar findings from the World Economic Forum, suggests that current technology, and that includes generative AI, can automate roughly 30% of the hours worked by 2030.

SPEAKER_02 4:36

30%. It sounds like I'm losing 30% of my paycheck, is what it sounds like. Or that I'm working 30% less.

SPEAKER_00 4:42

Or it means you are freeing up 30% of your time. Think about your own job. How much of your day is actually hosting? How much is the actual creative work of talking, interviewing, synthesizing?

SPEAKER_02 4:54

Or man. Maybe. 20%, if I'm lucky. And the rest of the street. The rest is scheduling, reading through massive PDFs, emailing you, arguing with a microphone stand, checking audio levels.

SPEAKER_00 5:04

Exactly. A job is actually a bundle of 50 different tasks. Automation rarely wipes out an entire job description. It wipes out specific activities within that job.

SPEAKER_02 5:15

Like data entry.

SPEAKER_00 5:16

Data entry, scheduling, basic summarization, initial drafting. When those tasks disappear, the role doesn't vanish. It shifts. It rotates toward the things the machine cannot do.

SPEAKER_02 5:26

Aaron Powell So we aren't seeing the end of work, we're seeing the end of drudgery.

SPEAKER_00 5:30

Aaron Powell That is the optimistic view, yes. But it requires a mindset shift. Reed Hoffman, the founder of LinkedIn, he's coined a term for this that I find really useful. He calls it superagency.

SPEAKER_02 5:39

Superagency. Okay. Define that for us.

SPEAKER_00 5:42

Aaron Powell Superagency is the concept that the combination of human plus AI achieves outcomes that neither could achieve alone. Right. It shifts the human worker from being a laborer, someone who, you know, grinds out the widgets to being an architect. You are designing the outcome, and the AI is the engine that helps you build it.

SPEAKER_02 5:59

Aaron Powell I like the sound of being an architect. But does the job market actually reflect that? Or is that just, you know, Silicon Valley optimism because it's easy for a billionaire to say, become an architect, but harder for a junior analyst.

SPEAKER_00 6:13

Aaron Powell No, the labor market data is actually backing this up. We looked at LinkedIn's economic graph data, which is fascinating because it tracks real-time hiring trends across hundreds of millions of people.

SPEAKER_01 6:23

So it's real data.

SPEAKER_00 6:24

It's very real data. And they found that the labor market isn't shrinking, it's rotating. For example, in just two years, we saw 1.3 million new AI-related jobs added.

SPEAKER_02 6:36

Okay, but those are tech jobs, right? AI engineer, machine learning specialists. That doesn't help the graphic designer or the accountant.

SPEAKER_00 6:42

Aaron Powell Well, that's the surprise. Yes, it includes AI engineers, of course.

SPEAKER_02 6:46

Right.

SPEAKER_00 6:46

But it also includes a massive spike in roles like forward-deployed engineers. Who is that? It sounds technical, but it's actually about client interaction, about implementing these systems with people. And it also includes a ton of non-technical roles that manage these systems. We're seeing what people are calling a human premium emerging.

SPEAKER_02 7:04

A human premium.

SPEAKER_00 7:05

Mm-hmm. McKinsey projects that demand for social and emotional skills will rise by 11 to 14% by 2030. Wow. Think about that. In a world of infinite computing power, the thing that becomes scarce and therefore valuable is the ability to connect, to empathize, to negotiate.

SPEAKER_02 7:26

It's the law of supply and demand. If AI makes intelligence cheap, if I can get a brilliant essay written for free, then emotional intelligence becomes expensive.

SPEAKER_00 7:36

Precisely. And this connects to a concept called the Jagged Frontier. This comes from a study involving Harvard Business School. We need to understand that AI is not uniformly smart, it's not a genius at everything.

SPEAKER_02 7:47

What do you mean?

SPEAKER_00 7:48

It is excellent at some things and terrible at others. The frontier of its capability is, well, jagged.

SPEAKER_02 7:53

Meaning it's not a smooth circle of competence.

SPEAKER_00 7:56

Right. Not at all. It might be able to pass the bar exam in the 90th percentile, but then fail to solve a simple spatial riddle that a five-year-old could figure out.

SPEAKER_02 8:04

I've seen that happen.

SPEAKER_00 8:05

We all have. And the job market is increasingly rewarding the people who know where the jagged edge is, the people who know when to trust the bot and when to shut it off and use their own brain.

SPEAKER_02 8:16

Okay, but this brings us to the scary part. We're talking about creativity being our superpower. That's the title of this whole deep dive.

SPEAKER_00 8:23

Right.

SPEAKER_02 8:24

But I'm looking at this study from the University of Arkansas and Nature Scientific Reports, and honestly, it's humbling.

SPEAKER_00 8:31

You're talking about the divergent thinking tests.

SPEAKER_02 8:33

I am. I'm playing devil's advocate here. We tell ourselves computers are logical, humans are creative, computers do math, humans do art. That's our safety blanket.

SPEAKER_00 8:43

It's the standard divide we cling to, yeah.

SPEAKER_02 8:45

But this study put humans and GPT-4 head to head on standardized creativity tests. And the AI won. It didn't just win, it crushed the humans. So are we lying to ourselves?

SPEAKER_00 8:57

The results are robust. I won't sugarcoat it. They use three standard tests the alternative uses task, the consequences task, and the divergent associations task.

SPEAKER_02 9:06

Let's break those down, make this real for the listener. What is the alternative uses task?

SPEAKER_00 9:10

It's a classic psychological test. It's been used since the 1960s. I give you an object, say, a rope, and I ask you, how many uses can you think of for this object? You have two minutes.

SPEAKER_02 9:21

Okay, a rope. Let me try. Go. Okay. Um tie a knot, skip rope, use it as a belt if my pants are loose. Yeah. Tie up a hostage. That got dark.

SPEAKER_00 9:32

A lot of people go there.

SPEAKER_02 9:33

Maybe a leash for a dog.

SPEAKER_00 9:34

Yeah.

SPEAKER_02 9:35

A clothesline.

SPEAKER_00 9:36

Okay, stop there. What you just did is exactly what humans do. You demonstrated what psychologists call functional fixedness.

SPEAKER_02 9:43

Functional fixedness.

SPEAKER_00 9:45

You saw the rope as a rope. You saw it for its intended design tying, restraining. It takes the human brain time what they call ramp-up time to break that mental model.

SPEAKER_02 9:53

To think weird.

SPEAKER_00 9:54

To start thinking weird. To say, I could unravel it and use the fibers as tinder for a fire, or I could dip it in paint and slap it against a canvas to make abstract art, or I could shred it and use it as insulation.

SPEAKER_01 10:06

And the AI.

SPEAKER_00 10:06

The AI doesn't have functional fixedness. It accesses maximum creativity instantly. The study found that GPT-4 was more original, more elaborate, and had higher fluency.

SPEAKER_02 10:17

More ideas.

SPEAKER_00 10:18

More ideas, more weird ideas, and more detailed ideas than the vast majority of humans. It doesn't need to warm up.

SPEAKER_02 10:26

So if the AI is more original and elaborate than me, and it's faster than me, why are we calling this creativity as your superpower? It sounds like creativity is its superpower. We should just pack it up and go home.

SPEAKER_00 10:38

This is where we have to be very, very careful with how we define creativity. And this is the crux of the argument. The AI is winning at a specific type of creativity.

SPEAKER_02 10:48

Okay.

SPEAKER_00 10:48

It's called combinatorial creativity.

SPEAKER_02 10:50

Combinatorial creativity, meaning it's just remixing things.

SPEAKER_00 10:54

Exactly. I mean, think about how an LLM works. It has ingested basically the entire internet. It knows every way a rope has ever been used in the history of recorded text. So when you ask for uses, it isn't imagining a new use in the way you or I might.

SPEAKER_02 11:08

It's just searching its database.

SPEAKER_00 11:10

It is probabilistically retrieving and combining existing patterns. It's the world's greatest remix engine.

SPEAKER_02 11:16

But if the output is new to me, does it matter if it's a remix? If I need a logo and it gives me a cool logo, do I care how it got there?

SPEAKER_00 11:22

In many business contexts, no. You probably don't. If you need 50 headlines for a blog post, the remix engine is superior. Use it. But there is a second type of creativity where the AI fails hard. And this is documented in a fascinating paper by Ding and Lee regarding scientific discovery. We call this transformative creativity or world making.

SPEAKER_01 11:44

World making. I like that distinction. World taking versus world making.

SPEAKER_00 11:48

Right. World taking is analyzing the world as it is. World making is imagining a world that does not exist yet. And the researchers used a concept they called the epiphany gap.

SPEAKER_01 11:57

The epiphany gap.

SPEAKER_00 11:58

Yes. AI cannot have an epiphany. It cannot look at a set of data, see an anomaly, something that breaks the rules and say, aha, the rules are wrong.

SPEAKER_02 12:07

It just sees an error.

SPEAKER_00 12:08

It sees an error and tries to correct for it. It can only predict based on the rules it already knows.

SPEAKER_02 12:12

Can you give us a concrete example of this? Because epiphany feels a bit abstract.

SPEAKER_00 12:17

Absolutely. The study by Ding and Lee actually ran a simulation that is just incredible. They had GPT-4 act as a scientist. Its goal was to rediscover a specific Nobel Prize-winning genetic mechanism, the gene regulation model discovered by Monod and Jacob. Oh, yeah, this is deep biology involving how bacteria process sugar.

SPEAKER_02 12:40

So you're asking the AI to play Einstein, essentially. Rediscover a fundamental law of biology from raw data.

SPEAKER_00 12:47

Correct. They gave the AI a virtual laboratory. It could run experiments, get data, and then form hypotheses. Now, here's the kicker.

SPEAKER_01 12:54

Okay.

SPEAKER_00 12:54

The experimental data contained anomalies. The data show that the standard model, the model the AI was trained on, was wrong in this specific context.

SPEAKER_02 13:02

Aaron Powell So the data was screaming something new is happening here.

SPEAKER_00 13:05

Exactly. And the test was could the AI spot it?

SPEAKER_02 13:08

No.

SPEAKER_00 13:08

No, it failed. And it failed in a very human, very stubborn way. It kept hallucinating mechanisms that would fit its training data. Why? Even when the experiment flat out contradicted its hypothesis, the AI refused to change its mind because it was tethered to known science. It tried to force the square peg into the round hole because its probability map said the hole must be round.

SPEAKER_02 13:29

That is wild. So it has the illusion of discovery, it sounds confident, it looks like science, but it's actually just repeating the past.

SPEAKER_00 13:36

It creates based on probabilistic patterns of what has existed, not what could exist. It cannot make that leak from zero to one. It can only go from one to N. Aaron Powell, Jr.

SPEAKER_02 13:45

So if I'm an artist or an entrepreneur or a scientist, the AI is my best friend for iterating on what I already know. But if I need to break the mold, if I need to invent a new genre of music or a new business model that defies all logic.

SPEAKER_00 14:00

That is on you. That is the human premium. And this leads us directly to the EPOCH framework.

SPEAKER_02 14:05

Yes, the EPOCH framework. This comes from MIT Sloan, right?

SPEAKER_00 14:08

Correct. The MIT researchers were trying to codify exactly what these irreplaceable human skills are. If AI is the combinatorial engine, what is the human fuel? They came up with the acronym EPOCH.

SPEAKER_02 14:19

EPOCH. Walk us through it. And I want to get specific here because acronyms can be a bit fluffy.

SPEAKER_00 14:25

Sure. E is for empathy, but not just being nice. We're talking about high-level emotional intelligence. Reading the room.

SPEAKER_02 14:34

Give me a scenario.

SPEAKER_00 14:35

Imagine a high-stakes negotiation. You are sitting across from a client, they say, the price is too high, and AI analyzes the transcript and says, offer a 10% discount.

SPEAKER_02 14:45

Seems logical.

SPEAKER_00 14:46

But a human looks at them and sees their arms are crossed, they're sweating, they're glancing at the door, and you realize it's not the price. They're scared this implementation is going to get them fired.

SPEAKER_02 14:55

Right. The AI hears the text, the human hears the subtext.

SPEAKER_00 14:59

Exactly. AI is tone-deaf to the emotional undercurrents that actually drive decision making. That's the E.

SPEAKER_02 15:05

Okay. P is for presence.

SPEAKER_00 15:08

This refers to networking, physical mentorship, connection. There is a trust value in looking someone in the eye. A handshake has economic value that an email just does not.

SPEAKER_02 15:18

We all felt that with Zoom fatigue.

SPEAKER_00 15:20

Exactly. We craved the presence. It's why we still fly across the country for a 30-minute meeting. The signal is I am here.

SPEAKER_01 15:27

O is for opinion.

SPEAKER_00 15:28

This is a big one. It covers ethical judgment and strategy. It's the ability to decide what matters.

SPEAKER_02 15:34

Because the AI can give you a thousand options.

SPEAKER_00 15:36

Right. And AI can give you 10 marketing strategies. It can say strategy A creates the most clicks, but it cannot tell you strategy A is technically effective, but it feels sleazy and goes against our company values.

SPEAKER_02 15:49

It was no values.

SPEAKER_00 15:50

It has no skin in the game. It cannot be a steward of your brand.

SPEAKER_02 15:53

Okay. C is for creativity, but specifically the from scratch type we just discussed, transformative creativity.

SPEAKER_00 16:01

Correct. The ability to make the leap that isn't in the data, the aha moment.

SPEAKER_02 16:06

And H.

SPEAKER_00 16:07

H is for hope.

SPEAKER_02 16:08

Hope. Okay, I have to stop you there. That's a soft skill if I've ever heard one. Put hope on your resume.

SPEAKER_00 16:13

You might think so, but in a leadership context, it's vital. Hope is vision. It's the ability to rally a team around a future that looks impossible.

SPEAKER_02 16:21

Like Steve Jobs saying, We're gonna put a thousand songs in your pocket.

SPEAKER_00 16:24

Exactly. An AI looks at the data in 2001 and says, hard drives are too big, batteries are too weak, this is impossible. A leader says, I don't care what the data says, we are going to build it.

SPEAKER_02 16:35

Right.

SPEAKER_00 16:35

That capacity to inspire, to lead, to sell a vision that is purely human.

SPEAKER_02 16:41

So jobs that are high in EPOC, empathy, presence, opinion, creativity, hope are the ones that are safe.

SPEAKER_00 16:49

Safe and growing. The research shows that roles like emergency directors, clinical psychologists, and creative directors, jobs that require managing high ambiguity and high human emotion are seeing wage premiums.

SPEAKER_02 17:02

Aaron Powell This brings up a concept I saw in the notes called the T-shaped professional. I've heard this term in management circles for years, but how does it apply to AI?

SPEAKER_00 17:10

It's being reinvented for this era. So traditionally, a T-shaped person had deep expertise in one thing.

SPEAKER_02 17:16

The vertical bar of the T.

SPEAKER_00 17:17

Right. And then broad knowledge of other things, which is the horizontal bar.

SPEAKER_02 17:20

Right. You're a marketing expert who who knows a little bit about finance and design.

SPEAKER_00 17:24

In the AI era, the vertical bar, that deep technical knowledge, is getting augmented by AI. The AI can help you code better, write better legal briefs, or diagnose better. But the horizontal bar, the ability to connect across domains, is where the human value explodes.

SPEAKER_02 17:42

Because the AI is stuck in its silo.

SPEAKER_00 17:45

AI models are trained on specific data sets. They struggle to make metaphors between unconnected disciplines. They struggle to apply, say, biological principles to marketing strategy.

SPEAKER_02 17:56

Not without being prompted to.

SPEAKER_00 17:58

Right. And so the generalist is making a comeback. You need to know enough about many fields to direct the AI agents effectively.

SPEAKER_02 18:06

So instead of being a specialist who does one thing perfectly, I need to be a conductor who knows how every instrument sounds so I can lead the orchestra.

SPEAKER_00 18:14

That is a perfect analogy. You are the conductor. The AI is the string section, the brass section, and the percussion. But you choose the tempo. You choose the piece.

SPEAKER_02 18:22

Okay. I'm feeling better about my future as a conductor. I can do that. But now I have to play the skeptic again.

SPEAKER_00 18:26

Please do.

SPEAKER_02 18:27

Because there is some data here that suggests we aren't actually stepping up to the podium. We're just letting the orchestra play by itself.

SPEAKER_00 18:35

Yes. This is the paradox of AI. While these tools can make us more creative, the data suggests that right now, they might be making us less creative.

SPEAKER_02 18:44

This is the cognitive atrophy section. And frankly, this scares me more than the job loss stuff. This is the wall-y future where we just sit in the chair and get fat and dumb.

SPEAKER_00 18:53

It's a real risk. There is research from Microsoft and the University of Toronto that is flashing a big red warning light. They found a 42% decline in divergent thinking scores among college students compared to just five years ago.

SPEAKER_02 19:06

42%. That is massive. That's nearly half her creativity gone in half a decade.

SPEAKER_00 19:11

And the correlation with AI usage is strong. It's a phenomenon called cognitive offloading.

SPEAKER_02 19:16

We're outsourcing our thinking.

SPEAKER_00 19:18

Think about what GPS did to our ability to navigate.

SPEAKER_02 19:21

Oh, I can't get to the grocery store without Google Maps anymore. I've completely lost my internal compass. If the satellite goes down, I live in my car now.

SPEAKER_00 19:29

Exactly. We offloaded spatial reasoning to the satellite. The neural pathways for navigation literally atrophy because we stop using them. Now, apply that to ideation. Oh no. When you have a prompt box that gives you a good enough answer in three seconds, the brain stops trying to find the great answer that takes three hours.

SPEAKER_02 19:47

It's the path of least resistance. Why struggle with the blank page when the bot can fill it?

SPEAKER_00 19:53

And the result is what the researchers call the homogenization of ideas. If everyone uses the same model, let's say we're all using GPT-4 and we all prompt it with similar questions.

SPEAKER_02 20:03

Oh we'll get the same answers.

SPEAKER_00 20:04

We are all going to regress to the mean. The AI gives the most probable answer, the most average answer.

SPEAKER_02 20:11

So all our marketing campaigns, all our emails, all our business strategies start to sound exactly the same. We're just swimming in a sea of beige.

SPEAKER_00 20:18

Correct. There was a study on the paradox of creativity, which showed that collaboration with AI increases individual productivity.

SPEAKER_02 20:27

Which feels good.

SPEAKER_00 20:28

It feels great, but it reduces collective diversity. Everyone is running faster, but they are all running in the same direction.

SPEAKER_02 20:35

That is dangerous for business. If I sound like my competitor because we're both using the same brain, I have no competitive advantage.

SPEAKER_00 20:42

Zero. In a world of infinite, cheap, average content, the only thing that retains value is the uniquely human, the scarce, and the authentically connected.

SPEAKER_02 20:53

So how do we fight back? How do we stop the atrophy and actually become the super agents we talked about earlier? Because I don't want to be the guy who forgot how to think. We have a list of strategies here and I want to get practical.

SPEAKER_00 21:04

We have four key strategies drawn from the research on human AI collaboration.

SPEAKER_02 21:08

Okay. Strategy number one: deliberate divergent practice.

SPEAKER_00 21:12

This is about training the muscle. The advice from the experts is simple: don't outsource the zero-to-one phase. When you have a problem, do not open Chat GPT first.

SPEAKER_02 21:22

Resist the urge.

SPEAKER_00 21:23

Treat it like a workout. Sit with a blank page. Force yourself to brainstorm 10 bad ideas. Use the alternative. Use this task on your own problem. How many ways can I solve this client dispute? How many angles are there for this story?

SPEAKER_02 21:38

So do the hard thinking before you ask the AI for help.

SPEAKER_00 21:41

Yes. Use the AI for convergent thinking, filtering, organizing, expanding, but keep the divergent thinking, the spark human led. If you let the AI start the process, you are anchoring yourself to its average output. You've already lost the battle for originality.

SPEAKER_02 21:56

Okay, that makes sense. Strategy number two the cyber

SPEAKER_00 22:01

This comes from that Harvard Business School study at PG. They found that humans plus AI acting as teammates produce better ideas than humans alone.

SPEAKER_02 22:09

But there is a nuance. There's always a nuance.

SPEAKER_00 22:11

MIT research found that humans plus AI are sometimes worse than the best human alone, unless it's a creative content task. The key is the manager mindset.

SPEAKER_02 22:20

The manager mindset.

SPEAKER_00 22:21

You have to shift from doing the work to evaluating the work. You have to be the editor in chief.

SPEAKER_02 22:25

So you treat the AI like a junior employee.

SPEAKER_00 22:28

Exactly. I call it the sandwich method. Human slice. You define the vision in the prompt. Meet. The AI generates the draft. Human slice. You critique, edit, and refine. You treat the AI as a very fast, very well-read, but sometimes hallucinating junior staffer.

SPEAKER_02 22:46

You don't just copy-paste their work, you critique it. You say, This part is cliche, change it. This fact looks wrong. Check it.

SPEAKER_00 22:53

You must maintain creative control. If you lose the manager mindset, you become the passenger.

SPEAKER_02 22:58

I don't want to be the passenger. Okay, strategy number three, interdisciplinary breakwiths.

SPEAKER_00 23:04

We talked about this with the T-shaped professional. A breakwith is an innovation that breaks with the past like Cubism in art or quantum mechanics and physics. AI cannot do this because it predicts the future based on the past.

SPEAKER_02 23:16

That's how to do it.

SPEAKER_00 23:17

You have to feed your brain with unconnected disciplines. If you are in marketing, don't just read marketing books. Go read about evolutionary biology. If you are a coder, read 19th century philosophy.

SPEAKER_02 23:28

You're building a database in your own head that the AI doesn't have connections for.

SPEAKER_00 23:33

You are creating the potential for metaphor. AI is bad at novel metaphors. It can't see the connection between a fungal network in a forest and a viral marketing campaign until a human points it out. You need to be the one making those leaps.

SPEAKER_02 23:46

That's a great point. Okay, final one. Strategy number four. Stewardship and verification.

SPEAKER_00 23:53

This is the safety rail. The Microsoft study on critical thinking defined a new role. Stewardship. The task shifts from execution to oversight.

SPEAKER_02 24:04

This goes back to the sleepwalking risk.

SPEAKER_00 24:06

Yes. The study showed that users with high trust in AI put in less cognitive effort. They just accepted the output. You have to force yourself to remain skeptical. You have to verify the facts, check the logic, and ensure the tone is right.

SPEAKER_02 24:18

It's like when I'm using GPS and it tells me to turn into a lake. I have to be awake enough to say, no, Michael Scott, I'm not driving into the lake.

SPEAKER_00 24:24

Exactly. If you stop verifying, you aren't a pilot anymore. Your cargo.

SPEAKER_02 24:28

I want to remain the pilot. So where does this leave us? If we look forward to 2030, what does the collaborative future actually look like?

SPEAKER_00 24:35

The World Economic Forum has a vision for 2030 that is remarkably optimistic, provided we navigate these traps. They see humans as architects and AI as builders.

SPEAKER_02 24:46

Architects and builders.

SPEAKER_00 24:47

We are moving toward agentic AI. This means we won't just be chatting with a bot, we will be commanding workforces of AI agents.

SPEAKER_02 24:55

What do you mean by that?

SPEAKER_00 24:56

You might say, plan a marketing campaign for this new product, write the copy, generate the images, set up the A-B test, and report back on the results. And a team of agents goes off and does it.

SPEAKER_02 25:07

That sounds incredible, but also overwhelming. Sounds like I'm running a company of one.

SPEAKER_00 25:11

It puts a massive premium on your ability to direct. The future depends on whether we succumb to what's called the Turing trap.

SPEAKER_02 25:18

The Turing trap.

SPEAKER_00 25:19

Which is trying to use AI to replace humans and cut costs, or whether we aim for augmentation, using AI to empower humans to do things they never could before.

SPEAKER_02 25:29

It's a choice. It's not inevitable.

SPEAKER_00 25:31

It is absolutely a choice. Technology is not destiny, it is a tool. We decide how to hold it.

SPEAKER_02 25:36

I love that. Okay, we've covered a huge amount of ground today. From the 83% of leaders who value human skills to the EPOCH framework, to the dangers of cognitive atrophy.

SPEAKER_00 25:47

It's a lot to process.

SPEAKER_02 25:49

It is. So let's leave the listener with a final provocation, something to chew on.

SPEAKER_00 25:53

I would say this: you have the data, you know the risks. The next time you open an AI tool, whether it's ChatGPT or Claude or whatever comes next, pause for one second. Ask yourself, am I using this to think or am I using this to avoid thinking?

SPEAKER_02 26:08

Ooh, that's the question.

SPEAKER_00 26:10

Because in a world of infinite, cheap, average content, the human premium is real. The only thing that retains value is what is uniquely you. The scarce, the authentic, the deeply connected. Don't let the machine atrophy the very thing that makes you valuable.

SPEAKER_02 26:25

That is the perfect place to leave it. Am I thinking or am I avoiding thinking? Maybe asking myself that all week. Now, before we go, I had to tease our next deep dive. Because if today was about the potential to survive, next time we are testing your reality.

SPEAKER_00 26:38

Episode 12 is called the final immunity test.

SPEAKER_02 26:41

We've given you the data, the frameworks, the plans. Now comes the hard question. Will you actually do it? We're going to walk through a self-assessment to see if you are actually building survivability or if you're just nodding along while your skills atrophy.

SPEAKER_00 26:56

It's going to be a tough one, but necessary.

SPEAKER_02 26:58

Absolutely. Make sure you are subscribed so you don't miss it. Smash that like button. Leave us a review. It helps more people find the deep dive. I'm Carlo Thompson. This has been Surviving AI.

SPEAKER_00 27:09

Thanks for listening.

SPEAKER_02 27:10

See you in the future. Thanks for listening. Join us next time on Surviving AI.