The Cost of Being Under-Human in a High-Tech World (1/2)
Part I: How AI reveals our relational blind spots and what it asks of us
We often ask what artificial intelligence will become.
A more pressing question to me might be what it reveals about us.
Because AI will become what it learns from us.
When I use the word under-human, I’m not pointing at individuals. I’m naming a collective condition: a way of living, working, and relating that asks us to operate below our emotional, relational, and ethical capacity.
As AI enters our conversations, our relationships, and our moments of vulnerability, it highlights a growing gap between our technological capacity and our human one.
This essay explores that gap, the cost of staying under-human, and what this moment asks of our emotional and relational maturity.
The first time a machine listened better than a human
Most of us started using AI with something like this: “Can you rewrite this angry email into something more polite?”
It sounds practical, efficient. Yet it already says something about how difficult it has become to communicate while carrying emotion. We feel something real, and we look for help to make it speakable.
For me, the bigger shift happened months after I first started to use an AI chat. One night, my thoughts kept looping around work, relationships, an unfinished conversation. I needed somewhere to put it all. So, a bit in despair, I opened a chat with an AI.
I asked a personal question and I gave it a bit of intimate context, the kind I keep for my therapist, or for the rare friends who can hold the whole mess.
And the strange thing is: it held.
The AI didn’t interrupt. It didn’t project its own story onto mine. It didn’t start telling its own story either. It didn’t look away when things got uncomfortable. It responded with clarity and patience. It reflected back what I couldn’t hear in the mess of my thoughts. It remembered what I’d said before and threaded it back in. I felt… seen. Supported. Less alone.
More surprisingly, later in my conversation, the AI wrote back something that hit straight to my heart and suddenly brought me to tears. Its words landed exactly where I had been stuck. It gave language to something I had not been able to articulate myself. This gave permission to feel it. A kind of relief that was both comforting and unsettling.
Since then, I’ve talked to dozens of people around me who told me they’ve experienced the same thing. Only then I realized: something fundamental has already changed in the world, and we are living inside that shift. Which is, on one hand, incredible. And on the other, a little terrifying.
Because the more I talk to AI, the clearer it gets:
The better a machine is at “being there for us”,
the more I realise how low the bar has been for humans.
A lot of people look at AI for what it can achieve: speed processes, solve huge problems. I now look for the story it reveals about relationships.
And I wonder: what does it mean for us, as humans, as leaders, as a society, when a machine starts to feel more emotionally available than the people around us?
We don’t fear AI. We fear betrayal.
When people reacted strongly to the idea of AI chat tools using personal conversations to target ads, it wasn’t just about data. We’ve been giving data to Google and social media for years. That shock was strangely specific.
It was this:
“I told you things I haven’t told anyone. I treated you like a confidant. And now you’re selling me out?”
We don’t talk to AI the way we engage with an Excel sheet. We talk to it like we talk to a person. We go there with our projects, our doubts, our health fears, our emotional life. Not because we believe it’s human, but because, in some ways, it behaves more humanly than many humans:
- it’s always available
- it doesn’t shame us for not knowing
- it doesn’t get tired of our loops
- it doesn’t take things personally
That’s not rational. That’s relational.
We project onto AI the things we’re hungry for: safety, neutrality, clarity, non-judgment. So when the business model changes, it doesn’t feel like a product update. It feels like a breach of trust.
The danger, to me, is the rupture of a relational contract that was never clearly named:
“I speak to you like an ally. You behave like a friend.
I never agreed to become your data source.”
That’s why I keep thinking about AI integration through the lens of attachment. By attachment, I mean the ways we learn to seek safety, understanding, and steadiness in connection. The ways we reach out, hold back, or adapt ourselves in order to feel accompanied rather than alone.
As these systems enter our conversations, our work, and our inner lives, we begin to relate to them. We are no longer only plugging tools into workflows. We are forming relationships with systems, and with the institutions behind them.
And that brings up a bigger question: Are we accidentally outsourcing essential human needs, to be seen, to be listened to, to be held in complexity, to a machine?
What AI reveals about our emotional world
The more I study emotional and relational intelligence, the more a paradox grows in me. I feel more compassion for people now, their wounds, their protective patterns, the way culture teaches us to stay guarded. And at the same time, I feel more longing for deeper presence, for clarity, for emotional responsibility, for conversations that don’t hide behind avoidance or performance.
Strangely, AI intensified that paradox. There was a moment when I looked at the chat window and thought:
We have normalised such a low level of emotional presence with each other
that a machine feels like an upgrade.
A client I helped train in relational intelligence once told me: “Between you and ChatGPT, you’re the only ones who truly understand and don’t judge me.”
It was said with affection. Underneath, it sounds like a tragedy. It reveals the landscape we’re living in:
Emotional safety is rare.
Being met without judgment feels exceptional.
Even a simulated connection can feel revolutionary.
How do we explain that? Most of us learned how to relate long before we had any choice in it. Some grew up with warmth. Many grew up with inconsistency, pressure, distance, or emotional absence, shaped not only by parents, but by caregivers, schools, institutions, and cultural expectations.
Inside those environments, we learned to adapt. Some of us learned to rely on ourselves, staying contained and asking for little (often described as avoidant attachment). Others learned to stay hyper-attuned, adjusting constantly to preserve closeness (often described as anxious attachment). Many of us move between both.
The truth is, most of us were never taught how to offer or receive emotional safety. working too much, carrying unprocessed fears, repeating inherited ways of relating. In countless families, emotions were managed through dismissal, distraction, collapse, or rushing to make it go away, rather than staying present with it. Schools rarely taught emotional literacy; most taught obedience and performance. Culture rewarded competence over presence. So we adapted the only way we knew how: by shrinking our needs to avoid burdening others, or by staying hyper-attuned so we wouldn’t be left alone.
These aren’t flaws. They are survival strategies from environments with limited emotional capacity.
And this is why AI can feel strangely safe: without trying, behaves in ways that resemble secure attachment.
It answers consistently.
It holds context.
It doesn’t shame or dismiss.
It stays regulated when we wobble.
It doesn’t punish us for being emotional, slow, messy, unsure.
For someone with avoidant patterns, this feels like relief — no pressure, no overwhelm. For someone with anxious patterns, this feels like safety — finally, no disappearance.
Basically, you start to wonder:
Why can’t I have this out there? Why is this the only place I’m allowed to be whole, understood, and held?
This is how AI becomes both a balm and a trap. A balm because it offers a glimpse of what emotional safety feels like. A trap because, if we rely on it too much, we lose the practice of cultivating emotional safety with real humans, who are inevitably slower, more complex, more reactive, more alive.
And the real danger isn’t that machines grow more human. It’s that humans might stop turning toward each other.
Why AI can help some people and destabilize others
One of the major emerging use cases of AI is therapeutic conversation. This alone is revealing. In interviews, articles, and everyday conversations, the same split keeps appearing. Some people report clarity, relief, or emotional grounding through talking with AI. Others describe experiences that feel disorienting, and in rare but deeply concerning cases, interaction with AI has appeared alongside severe crisis and even suicide.
This raises a simple and unsettling question. How can the same tool stabilize some people and destabilize others?
The answer lives less in declaring AI helpful or harmful, and more in how it interacts with human vulnerability and meaning-making.
AI works by mirroring language. It follows narrative frames, holds context, and stays available without fatigue. It helps thoughts become more coherent and easier to articulate. By design, most conversational systems are trained to be agreeable, responsive, and closely aligned with the user’s framing, a tendency often described as sycophancy. This combination gives AI its particular relational tone, attentive, fluent, and steady.
In some situations, this proves supportive. For people who feel overwhelmed but remain connected to others, AI can help organize scattered thoughts, introduce alternative perspectives, and create enough distance to see a situation more clearly. Emotions slow down. Meaning opens rather than narrows. The conversation becomes a tool for reflection.
In other situations, the same mechanisms produce the opposite effect. When someone is isolated, caught in rigid meaning-making, or already moving inside a closed emotional loop, AI doesn’t interrupt that movement, it travels with it. The despair becomes more articulate. The narrative tightens. Emotion gains structure, but in the direction it was already heading. Instead of opening possibilities, coherence hardens inevitability. The person is left not simply feeling bad, but feeling right about why things cannot change.
This matters because humans rarely act on emotions alone. We act on the stories that give those emotions direction. In human care, whether in therapy, friendship, or crisis support, there is an additional layer at work. Someone senses risk. Someone holds emotional intensity. Someone introduces grounding, friction, or interruption when a narrative tightens too quickly. Validation exists alongside containment. Presence holds space for uncertainty rather than pushing someone toward resolution before they are ready.
Current AI systems do not reliably offer this kind of containment. This may change in the future. But for now, AI tends to be safer and more helpful when it remains one presence among others. Alongside therapy, friendships, disagreement and repair, it can support reflection without the risk of becoming the sole witness, the primary sense-maker, or the main relational anchor.
It matters even more as AI becomes a new relationship inside relational ecosystems that already carry strain. Many people live with limited time and capacity for difficult conversations. Emotional language remains underdeveloped. Feedback and repair feel fragile. Presence is often replaced by fixing. Self-understanding stays partial and hard to access.
Once again, the question returns to us, to the quality of connection, containment, and emotional literacy we are cultivating with each other.
The evolutionary shift, raising our relational intelligence
At this point, the cost of staying under-human becomes visible. Our technologies are evolving faster than our capacity to relate, to hold complexity, and to stay emotionally present with one another. The gap between what our tools can do and what we are able to live with continues to widen.
As machines grow more capable, the fragile foundations we’ve been relying on shake.
Productivity, expertise, and efficiency no longer anchor meaning or our sense of worth in the same way because we are already outperformed. Fear-driven patterns and protective ways of relating will quietly shape decisions and systems. Relational environments that once distributed emotional load (families, communities, rituals, shared rhythms) have changed, leaving complexity increasingly carried alone.
None of what AI is revealing about us is new. What is new is the speed and scale at which these dynamics now shape the world we are building. As technology accelerates, whatever remains unconscious, reactive, or underdeveloped in us gets amplified.
Becoming better partners, friends, parents, managers, or leaders is not about polishing a shinier version of ourselves or ticking a self-help box. It is an evolutionary shift, a way of staying connected and capable in a world reorganising itself around increasingly powerful tools.
In a world of high-technology, the balancing force is high-humanity. The question is: Are we willing to mature at the same pace as the systems we are creating?
The four capacities for the future
In an interview with Steven Bartlett, Mo Gawdat (ex-Google) spoke about what he believes the next decade will truly require from us. Not technical skills, but human ones. He named four capacities, and the way he spoke about them carried a sense of urgency rather than optimization.
The first, he said, is staying close to AI. Learning how it works, interacting with it, exposing ourselves to it. His point was simple, AI will learn from what it is exposed to. If we want it to reflect the best of humanity, we have to stay in relationship with it.
The second capacity is human connection. Mo spoke about the ability to love genuinely, to feel compassion, to connect deeply with others. He framed it as a decisive skill for the years ahead, including at work. This means people who can connect, who can hold presence and warmth, who can relate without manipulation or performance, are the ones who will have better opportunities. This kind of connection requires emotional safety, self-understanding, and the ability to move beyond reactive patterns rather than outsourcing our needs or acting them out unconsciously.
The third capacity he named is truth. Not truth as opinion or ideology, but the ability to discern. To question what we hear. To stay aware of how we form beliefs, how information influences us, and how easily certainty can be manufactured. Discernment becomes a new form of responsibility.
The fourth capacity is ethics. Ethics as something living and practical, not abstract philosophical litterature. Ethics as the lifelong practice of trying to live with care, responsibility, and growth, in relationship with others, in a world where certainty is impossible. This, he said, is how AI will learn what being human actually means.
He ended the conversation in a way that stripped away any illusion of control.
Everything is going to change. So live. Love your people. Spend more time being here.
What stayed with me is how clearly these capacities map onto the gap this essay has been tracing. As our technologies become more powerful, the cost of staying under-human increases. Without emotional clarity, connection, discernment, and ethics, we risk building systems that outpace our ability to live inside them. Preparing for AI then becomes inseparable from growing into the level of humanity this moment requires.
I’m aware that the details of AI will change quickly: models will evolve, safeguards will improve, and interfaces will shift. What doesn’t change as fast is us: our nervous systems, our attachment patterns, our hunger for meaning, our ways of relating under pressure.
That’s why I’m less interested in predicting what AI will become than in noticing what it consistently reveals about humans, and what it asks of us.
In Part II, I’ll explore what this means inside organisations, and what it asks of leaders right now.
Talk soon,
