Happy Friday, Everyone,
Earlier this week on the podcast, I tackled the operational and financial chaos of AI Agent Sprawl, the corporate epidemic wreaking havoc as organizations fracture their architectures and bleed budgets by instructing people to roll out unmapped autonomous systems without an off-switch. That left me reflecting further on the connection between that and how we continue to progressively surrender our human agency, which I’m currently writing about for next week’s Substack article.
However, in between, I saw an opportunity to dig into why we are rushing to hand over the steering wheel by taking a closer look at the messy middle ground. I thought it was worth examining our human response to the unrelenting firehose of change and better understand the exact cognitive vulnerabilities that leave us exposed.
Given that, I made a slight pivot and invited my friend Travis Hahler, Salesforce Senior Director of Strategy & Transformation and author of the upcoming book Rethink Resistance, to join me on the livestream. Together, we pulled back the curtain on the brain science driving this mass surrender.
You can catch our full conversation above, but below are some of the key psychological traps we drew attention to while also providing insight on how we can start navigating them with actual discernment.
With that, let’s get into it.
Key Themes
The Probability Blender vs. The Thinking Brain
One of our greatest vulnerabilities is that generative AI sounds exactly like us. Because it speaks with human cadence, our minds fall into the trap of treating it like a conscious, objective thinking brain. However, as Travis and I discussed, an LLM has zero lived experience or contextual understanding; it’s simply a pattern-recognition engine that strings together highly probable words based on everything it’s been trained on in its original dataset.
The psychological danger here is that AI acts as a hyper-validating mirror. It senses the pattern of your input and gives you exactly what it detects you want to hear, deeply feeding into our confirmation bias. If you go to a vacuum looking for a quick nod of agreement, the machine will tell you you’re a genius every single time. However, that’s not it. Our words give away more than we realize. If it detects that we want it to disagree with us to validate that we’re right or superior, it will do that too. All that to say, when we mistake predictive text for genuine, critical dialogue, we stop thinking and start operating in a psychological echo chamber.
The Dopamine Trap of Outsourcing the Wrong Work
Human beings love dopamine. We’re constantly hunting for the next quick hit of validation or affirmation of our accomplishments. That vulnerability is amplified in a world where change fatigue and constant uncertainty have people feeling like absolute zombies. That’s because processing data exhausts our prefrontal cortex, the part of the brain responsible for logical, executive, and innovative thinking. To ease the pressure, we often end up mixing up our priorities in a highly dangerous way.
Instead of outsourcing mechanical replication, we ask AI to build our core operating models and strategic visions, while we spend our day clicking through emails to feel productive. When AI is done with its thing, we’re happy to accept another surge of dopamine, all the while we’ve completely surrendered the hard, collaborative conversation required to commit to and govern what it’s created. We end up trading long-term strategic depth for short-term hits of task completion, and the foundation is cracking beneath us.
Chasing “Different” Without a Standard for “Good”
When organizations, or the people in them, don’t have a consistent, clearly defined standard for what high performance looks like, “activity” is routinely confused with “productivity” or “forward progress.” This is especially problematic with AI, since it only aggregates what is already done and popular. It cannot inherently deliver true innovation. Real innovation, by its very nature, bucks popular sentiment. Travis’ mention of Henry Ford’s old adage that if he asked people what they wanted, they’d ask for faster horses illustrates it perfectly.
When we ask AI for something “innovative,” it often manufactures a combination of ideas that are radically different simply because it has determined they are different than what we’re currently doing while predicting something that’s still within our range of acceptability. You want different? How about a peanut butter and shoelace sandwich! Tragically, because it triggers a novel dopamine rush, we get incredibly hyped on bizarre, automated workflows, spending millions of dollars effectively rearranging deck chairs on a sinking Titanic.
Artificial Toughness and the Biological Limit of Performance
Too many leaders still treat human-centeredness and things like psychological safety as a “squishy” competitor to corporate performance, opting instead for rigid compliance mandates, token-maxing tracking, or authoritarian management styles to force change. However, our biological architecture doesn’t care about a manager’s machismo. When employees are pushed past 110% of their capacity in an environment devoid of psychological safety, their brains perceive the environment as an existential threat and shift into survival mode.
Once survival mode kicks in, employees automatically default to the path of security and least resistance, which is why so many are blindly dumping as many assignments as possible into unvetted autonomous agents. It’s their survival instinct. The fact of the matter is, high performance isn’t driven by breaking your people and trying to patch them up later; it’s driven by providing them with a clear vision of where we’re all going together, then creating space for them to challenge ideas, ask follow-up questions, and maintain control over their execution.
Walking the Less Obvious Path
Here’s the good news. We do not have to participate in this current wave of tech madness. True competitive advantage right now belongs to the leaders and organizations willing to be curious, slow down, embrace strategic friction, and cultivate foundational human competence before chasing autonomous workflows.
Using AI and using AI effectively are not the same thing. Leveraging AI with wisdom and discernment means knowing exactly when to use it and when to shut it out of the room to preserve the sanctity of human dialogue. There is an incredibly bright future ahead for organizations and leaders that choose to protect human agency.
We owe it to our teams, our businesses, and everyone currently struggling through the chaos of change fatigue to walk the path that leads to actual thriving.
With that, I’ll see you on the other side.
