There's a paper making rounds in evolutionary biology circles that most business leaders will never read. Published in Proceedings of the National Academy of Sciences in Fall 2025, it asks a question that sounds like science fiction:
Could humans and AI become a new evolutionary individual?
The authors aren't futurists. They're evolutionary biologists from the Max Planck Institute for Evolutionary Biology and the University of Montpellier. And their argument is grounded in how life has always worked.
Consider what you already are.
Your cells contain two distinct organisms: the nucleus and mitochondria, each with its own genome. Mitochondria were once free-living bacteria. Somewhere around two billion years ago, one cell engulfed another, and instead of digestion, there was integration. The merger became the foundation of all complex life.
Your body hosts microbiomes, in your gut, on your skin, across every mucous surface. These aren't passengers. They're inherited ecosystems, passed from mother to child, that shape your immune function, your metabolism, possibly your mood.
We have always been more than we thought we were. The boundary of "self" has always been permeable, held together by something other than isolation.
So when researchers ask whether humans and AI might form a new composite organism, they're describing a pattern. Life moves toward integration. The question is what kind.
What changes when AI changes us
A study published this year in Computers in Human Behavior tested what happens to human cognition when we work with AI. Researchers gave participants logical reasoning problems from the LSAT. One group worked with ChatGPT. A control group worked alone.
The AI group performed better, about three points higher on average.
But something else happened. When asked to estimate their own performance, the AI-assisted group showed significant overconfidence. They believed they'd answered 17 out of 20 correctly. The actual number was closer to 13.
More striking: the pattern of overconfidence changed. Without AI, the data shows the classic Dunning-Kruger curve, low performers vastly overestimate themselves, high performers don't. With AI, that curve vanished. Overestimation became uniform. Everyone, novice and expert alike, inflated their scores by similar amounts.
The technology didn't just augment performance. It restructured how people perceive their own capacity.
We're losing the signal that tells us where we're weak. The inner voice that says slow down, you don't know this yet goes quiet.
Two trends, one question
Read all of this together.
Integration is accelerating. The composite human-AI unit outperforms in ways that get rewarded, careers, resources, access.
And the integration itself is dulling something essential: our ability to hear ourselves clearly. To know what we know and don't know. To pause before the cut.
If we can't hear that signal, what else might we stop hearing?
Major evolutionary transitions need not necessarily be adversarial. Sometimes separate entities merge because cooperation creates possibilities neither could achieve alone, a principle relied on heavily within our Supply Chain of the Future collaborative work. But ancient bacteria didn't have a choice. We do. The merger between humans and AI won't be inhibited by physics or chemistry but rather shaped by the systems we design, the boundaries we establish, and the values we embed in our infrastructure.
What gets passed down
This isn't an argument against AI. I've spent thirty years implementing technology systems, from dial-up to gigabit, from punch cards to machine learning. As Principal at TOTEM and Co-Founder of COSSAF, the Collaboratory for Open Software and Systems in Ag and Food, we're building AI infrastructure right now with organizations across the fresh produce industry. I believe in what these tools can do.
But I also learned the trade from my father. He was an electrician. I apprenticed with him as a kid, running wire through crawl spaces, watching him solve problems in real time. He taught me something I didn't have language for until much later: the quality of your work is inseparable from the quality of your attention.
He would pause before cutting. Check twice. Not because he was slow, he was fast when it mattered, but because rushing through the wrong moment costs more than the seconds you save.
That pause wasn't inefficiency. It was integrity. It was listening before acting.
And it was inheritance. Something passed down. Something I didn't invent and can't afford to lose.
The choice
The question isn't whether we'll integrate with AI. That's already happening. The question is whether we'll do it in a way that honors what we've been given, the capacity for attention, for reflection, for hearing the still signal underneath the noise.
Or whether we'll barrel forward, uniformly confident, unable to hear what we've stopped listening to.
This series is about the practical work of AI integration in complex organizations. Strategy, methodology, proof of what works.
But underneath the tactics is a deeper question: how do we build systems that preserve what matters most about being human?
I don't think the answer is resistance. I think the answer is design.
We can build systems that keep humans in the loop not as a governance checkbox, but as the point of connection. We can create architectures where the pause isn't inefficiency, it's where the listening happens. We can choose integration that empowers rather than absorbs.
The mitochondria didn't choose the merger. The microbiome doesn't either.
We were made for something different. We can choose.