When was the last time you memorized a phone number? For most of us, that answer is buried somewhere in the ’90s, alongside dial-up internet and pagers. What used to be an automatic skill is now outsourced to our devices, and our brains have literally rewired in response. That’s the starting point of Dr. Poppy Crum’s research and one of the most eye-opening reminders from her recent talk at Proofpoint Protect 25, moderated by Proofpoint Chief Strategy Officer Ryan Kalember (pictured above): Technology doesn’t just change the tools we use. It changes us.
Crum, a neuroscientist, technologist, and futurist who has led innovation at Dolby Laboratories, Trimble, and now Goby Technologies, is one of the world’s sharpest voices on how human evolution and AI are colliding. Her work sits at the intersection of physiology, technology, and cognition, and she’s making the case that AI isn’t just transforming industries. It’s transforming our neural pathways, our ability to think, and even our vulnerability to threats like cyberattacks.
The Brain on AI: What’s Really Changing
Crum shared a striking example: London cab drivers once had to memorize every street in the city before they could earn their license. That training physically reshaped their hippocampus, giving them more gray matter than the average person. But today, with Google Maps in every passenger’s hand, that brain boost has vanished. Outsourcing memory changed their brains.
AI is now taking this outsourcing to a new level. A recent MIT study compared students writing papers with and without ChatGPT. The difference wasn’t just in output—it was in brain activity. Students who leaned on AI showed less engagement in the prefrontal cortex and working memory. They produced content, but they failed to build the mental scaffolding that allows ideas to connect, deepen, and transfer into long-term learning.
As Crum put it: When you let the agent think for you, you risk losing the neural wiring that makes critical thinking possible.
Cognitive Load: The Hidden Cost of Efficiency
Crum breaks cognitive load into three categories.
- Intrinsic load is the effort required to actually learn or solve something difficult, like understanding a new security architecture.
- Extrinsic load includes distractions and noise that make the task harder than it should be, like acronyms, jargon, or constant alerts.
- Germane load is the “good” load, where your brain builds connections, strengthens pathways, and learns how to learn.
AI reduces intrinsic and extrinsic load, which sounds great. But it also threatens germane load. If your brain never does the heavy lifting, it never grows stronger. That creates a dangerous trade-off for both individuals and organizations: speed at the expense of resilience.
The Cybersecurity Parallel
This isn’t just theoretical. In cybersecurity, attackers already exploit cognitive overload, using urgency, panic, and distraction to trick people into clicking. Now, with AI capable of hyper-personalization, spear-phishing can be timed for the exact moment someone is stressed, tired, or distracted.
Crum warned that this cuts both ways. Defenders need to leverage AI to anticipate when humans are vulnerable and step in before a mistake happens. That means using context-aware tools that reduce extrinsic noise while preserving the human’s ability to detect anomalies and exercise judgment.
Put simply: AI can either sharpen our defenses or dull them, depending on how we train both machines and people.
The Human-AI Future: Managers, Not Passengers
Crum doesn’t argue against using AI; far from it. She argues for conscious adoption. Her advice is clear:
- Don’t just outsource. If AI is doing something for you, find ways to still exercise the underlying cognitive skill elsewhere.
- Stay in control. Employees must be able to reverse-engineer an AI’s output, not just accept it at face value.
- Train for difference. A junior employee who relies heavily on AI will think differently than a senior leader who built knowledge the “hard way.” Organizations need to close that gap, not ignore it.
- Make the brain work. Neuroplasticity thrives on effort. If your brain isn’t working, it’s weakening.
As she put it bluntly: “If you’re not working your brain, that’s not good. You want your brain to be working.”
Why It Matters
Technology has always reshaped human behavior. Cars changed our cities, phones changed our attention spans, and now AI is changing how we think. But unlike past innovations, AI’s impact is deeply personal: It’s altering the very wiring of our minds.
For businesses, this means the AI conversation can’t stop at efficiency gains. Leaders must ask: What does this do to my people’s ability to think, decide, and innovate? Because in the future of AI-driven work, competitive advantage won’t come from outsourcing thought. It will come from balancing human creativity with machine speed.
And that balance? It starts in the brain.
