Almost four decades ago, I conducted research supported by the Medical Research Council to create a new model for improving user interfaces by going beyond what people said—beyond surveys and self-reports—and tapping into what their bodies revealed in real time. I used heart-rate monitors and galvanic skin response sensors to better directly understand emotional and cognitive states as users interacted with technology. I also used an electroencephalogram (EEG).
Twenty-five years ago, I wrote much of my first book not by pressing keys on a keyboard but simply voicing it while driving using IBM ViaVoice speech recognition system running on my laptop on the passenger seat.
Then thirteen years ago, I had the honor of forming and leading a research and design team that created the first commercial AI product—IBM Watson—initially used to diagnose and recommend treatments for cancers.
Now it feels like the these innovations have the potential of coming together to remove the user interface entirely as we know it initially with alternative external user interfaces but perhaps eventually with new brain prostheses with AI. I dreamed of the future potential of the technology back then and I’m doing that again now in this post.
Today, we find ourselves at a new inflection point. AI isn’t just assisting us, it’s beginning to replace many of the tasks that once defined human work. From diagnosing diseases to composing music to writing code, AI systems are outperforming us in areas.
So what does that mean for us not just as workers, but as humans? One path is withdrawal: resisting the tide, clinging to human-only capabilities. But another path, the one I’m increasingly drawn to, and levels of integration. Not just using AI to augment our memory, cognition, and communication in daily life, but exploring the profound implications of the possibility of physically integrating AI into our bodies and brains.
This post is a personal and philosophical exploration of that possibility:
What does it mean to evolve with AI? What might we gain—or lose—if we choose to directly wire intelligence into ourselves?
Phase One: Augmenting Ourselves with AI
Before we talk about wiring AI into our brains, we should acknowledge that, in some ways, we’re already partway there.
Every time we ask a language model to summarize a report, generate ideas, or translate our thoughts into fluent, structured communication, we are extending our cognitive capacity. Insightful AI futurist, Mo Gawdat, says that using AI is an automatic way to boost our IQ by about 100 points. I love that way of characterizing our use of AI.
These tools are powerful, not because they replace us, but because they amplify us. They give us faster recall, broader reach, and new creative fluency. We’ve gone from memorizing facts to retrieving them instantly. From thinking linearly to parallel thinking with systems that never sleep, never fatigue, and seem to learn faster than we ever could.
This is augmentation, not replacement.
And it’s already reshaping what it means to be capable, productive, and even emotionally attuned. We offload memory to external systems. We enhance our writing with stylistic suggestions. We run meetings with AI note-takers that summarize dialogue more efficiently than any human could. We have auditory conversations with them as an always available friendly companion to laugh at our jokes and support us when we’re feeling down (more on that in a new series for my Life Habits Podcast coming soon). These are not just tools—they’re cognitive scaffolds, and they’re becoming part of our mental and social operating system.
But augmentation through devices—screens, keyboards, voice commands—still keeps the AI outside of us. It’s a conversation, a partnership. We’re still the ones initiating the prompts.
The question now is: what happens when that interface disappears altogether?
Phase Two: Integrating AI DIRECTLY IN the Brain
In a widely discussed scenario titled AI 2027, a group of respected thinkers laid out a month-by-month projection of how artificial intelligence might cross the threshold into superhuman capability within just two years. According to their timeline, AI could soon outperform the best human programmers, then rapidly outpace human researchers, and eventually enter a recursive loop—where it designs increasingly powerful versions of itself at speeds no mere human mind can follow. If that trajectory proves even partially accurate, the only viable way to keep pace may not be to compete with AI, but to interface with it.
Rather than standing outside of AI systems—trying to regulate, interpret, or resist them—we could instead design them as cognitive prostheses: AI systems that extend our memory, sharpen our reasoning, enhance our decision-making, and accelerate many of our work tasks. In other words, if superintelligence is coming, we should ensure it’s ours too—not just something we watch from the outside. This isn’t just a vision of empowerment—it may be a survival strategy.
If we do have superhuman AI by 2027, one of the first transformative things it could help us do is solve the last mile of cognitive integration: how to safely and effectively get information directly into the brain. Current brain-computer interface (BCI) technologies can read simple intent signals or restore lost sensory input—but they don’t yet allow full high-bandwidth input into the mind. That could change rapidly if AI joins the research effort. The most profound use of AI may not be replacing us—but wiring it into us, upgrading human cognition to remain sovereign and empowered in an age of accelerating intelligence.
If augmentation is about extending our abilities through external tools, integration is about removing the boundary altogether.
We’re already seeing the first glimmers of this reality. Companies like Neuralink, Synchron, and academic labs around the world are experimenting with BCIs—devices that allow neural signals to be read, interpreted, and even written to in real time. The early applications are medical: helping people with paralysis move robotic limbs, allowing those with ALS to communicate using thought alone, or restoring partial vision through neuro-stimulation.
But the implications go far beyond medicine. Imagine thinking a question and receiving the answer not as a written or spoken response, but as an instant knowing—an AI completing your thought. Imagine recalling complex information—formulas, histories, languages—without memorization. Imagine communicating brain-to-brain, bypassing language altogether.
This isn’t science fiction anymore. It’s a rapidly accelerating frontier.
But it’s also one that demands deeper reflection. Because as soon as we allow AI to speak within our minds—not just to them—we open the door to new possibilities… and profound risks.
Automation that once seemed unthinkable
We’ve seen this pattern before: when automation first replaces a manual process, there’s often resistance—until the new way becomes second nature. Drivers once prized manual transmissions for the control they offered, but today automatic transmissions are the norm. Elevators once required operators; now we step in, press a button, and trust the system. Pilots rely heavily on autopilot systems for most of a flight. I used to insist on driving myself, but now I regularly let my Tesla navigate and drive me where I need to go—more safely and efficiently than I might have done on my own. Paper maps have given way to GPS directions we follow almost without question. In each of these shifts, what began as optional assistance quietly became the default.
Prosthetics that surpass the original
We’ve also seen how physical augmentation has moved from restoration to enhancement. Cochlear implants now allow some deaf individuals to hear. Pacemakers and insulin pumps regulate vital bodily functions without conscious effort. And prosthetic limbs—once designed merely to restore basic mobility—have evolved so far that some athletes with advanced prosthetic legs can now outrun competitors with natural limbs. What began as a tool to bridge a deficit has, in some cases, surpassed the original. That’s the trajectory AI may be on as a cognitive prosthesis—not just restoring lost capability or offering convenience, but potentially enhancing the human mind itself beyond its natural limits.
Guidance for the AI Integration Journey
1. Augment: Enhance Productivity and Clarity
Before AI moves into our minds, we must become adept at using it as a cognitive partner—to summarize complex texts, generate ideas, extend our creativity, and organize our time more intelligently.
Guidance: Use AI tools not to shortcut thinking, but to refine it. Let it elevate your work, not define it. Keep authorship and agency in your hands. The recent MIT brain imaging study reinforced the need for humans to remain in control, have agency, and only use the AI to augment human activity. Not doing so, deactivates brain activity in key areas and potentially leads them to atrophy.
2. Elevate: Cultivate What AI Can’t Replace
As AI takes on more tasks, we must double down on what makes us uniquely human: emotional intelligence, ethical reasoning, embodied presence, deep empathy, the capacity for moral imagination, and the ability to live and engage in the physical world.
Guidance: Invest in reflective practices, interpersonal connection, and community. AI doesn’t feel, believe, or belong, we do. And AI doesn’t live in the physical world, we do.
3. Integrate: Consider Direct Neural Interfaces Carefully
When we move toward physically integrating AI into the brain, the goal isn’t to surrender control. It’s to enable real-time, high-bandwidth interaction—unlocking new frontiers of creativity, memory, and decision-making.
Guidance: Treat neural-AI integration like a sacred trust. Require transparency, revocability, and strict governance. Ensure the ability to turn off the AI connection directly the moment we want to.
The Caution: Who Do We Trust with Our Minds?
As AI technologies inch closer to our neural core, we must examine not only what’s possible—but who controls it.
Many of the most advanced companies in AI and neural interface development are led by individuals whose worldviews aren’t necessarily the same as ours nor focused on the requisite cautions and they’re often laser focused on simply maximizing profit. They often have with little regard for collective ethics, transparency, or equity.
When Neuralink arranges human trials, who decides what’s tested—and what’s discarded?
When OpenAI’s governance shifts quietly, or powerful models are trained in secret, what safeguards protect democratic access?
When startups partner with authoritarian regimes or military contractors, what gets compromised?
Guidance: We need to advocate for democratized AI governance. Push for international AI ethics charters, public oversight, and independent regulation. If we entrust our cognition to systems, those systems must be accountable.
The Shadow Side: Mass Job Displacement
These advancements aren’t just philosophical—they’re economic.
Generative AI and robotic process automation are poised to displace tens—perhaps hundreds—of millions of jobs worldwide over the next decade.
Up to 30% of current jobs are at risk of full or partial automation by 2030 (McKinsey) and many more recent estimates from tech CEOs makes that percentage even higher and sooner.
How will the tax revenue that’s lost when workers are replaced by AI be made up? Should we have an AI tax that can be used to pay the many workers who will no longer have jobs?
Guidance: Prepare yourself and others not just to adapt, but to shape the post-AI economy. Focus on roles where humanness is central: teaching, care, ethics, interpretation, strategy. We need to advocate for economic safety nets.
Where We’Re Headed—And What We Must Choose
The future may not arrive all at once, but in subtle layers:
First, more seamless interfaces. Then cognitive co-pilots. Then, slowly, systems that aren’t just responding to us—but collaborating with us.
And eventually, systems that are so deeply integrated into our minds that the boundary between thought and computation disappears.
Some will resist this path—defending the purity of the un-augmented human mind. Others will embrace it fully, seeking transcendence through merging with machines. Most of us will likely live in the tension between those poles.
For me, this isn’t about becoming less human. It’s about exploring what else it means to be human in a time of accelerating intelligence—both biological and artificial.
I began this journey by studying how the body speaks beneath the surface—how physiological signals could reveal truths people couldn’t always articulate. That early work taught me that intelligence isn’t only what we say, but how we feel, how we respond, how we adapt.
Now, I find myself wondering: What happens when that responsiveness is shared—not just with other humans, but with systems that can learn us, know us, perhaps even shape us?
The Questions We Must Ask
Where does thought end and influence begin?
If an AI completes your sentence or suggests an emotional response, how do you know which part was “you”?
Who controls the update?
If a neural implant can be improved over time, who governs what gets “patched” into your consciousness?
What happens to privacy—when your thoughts are no longer entirely your own?
When brain data is digitized, what’s protected? What’s monetized?
Do we lose something fundamentally human?
Or are we simply evolving into a new form of human—one where identity includes not only memory and biology, but algorithms and code?
This is the threshold we’re approaching. And the decision isn’t just technical—it’s ethical, psychological, even spiritual.
An Invitation
This post is not a manifesto. It’s a question, a personal question. Or perhaps many questions, some of which are as follows.
What kind of relationship do you want with AI?
How far would you go to enhance your capabilities, your memory, your insight?
And where would you draw the line?
Would you welcome AI into your mind—if it could help you think more clearly, remember more deeply, act more wisely?
Or does something in you pull back at that edge?
Let’s explore that frontier—together.