🐘 The Elephant in the Room Series
The Matrix taught us to fear machines taking control. What if the more immediate handoff is the one we make by invitation?
We spend a lot of time talking about what AI might eventually take from us.
Our jobs. Our creativity. Our ability to think. Our control.
But there is another question I hear far less often:
What are we willingly handing over?
Not only our data, workflows, or the next task on a list.
Our attention. Our judgment. Our ability to listen. Our creative confidence. The practices through which we learn what we think, what we value, and what we want to stand behind.
That question brought back a memory from college.
Sorry, Dave
When I was studying Information and Computer Science at UC Irvine, I was also, occasionally, a bit of a prankster.
One day, a fellow student named Dave left his computer accessible. A few of us changed the sign-in interface so that whenever he tried to log in, a message appeared:
“Sorry, Dave. I can’t do that.”
If you know 2001: A Space Odyssey, you know the reference: HAL 9000, the calm and unsettling computer that appears to overrule the humans who created it.
Of course, Dave’s computer had not developed an opinion. It had not become autonomous. It was not refusing him.
We had changed the interface.
But that was exactly why the prank worked. For a moment, the machine appeared to be exercising agency over a person.
In reality, the agency was still ours.
Decades later, that silly college story feels unexpectedly relevant. Not because today’s generative AI is merely an illusion. It clearly does far more than display a cleverly timed message. It can write, create images and music, summarize, recommend, analyze, critique, and, through connected tools, increasingly act on our behalf.
The point is different: we remain intensely focused on the question of what agency the machine has.
We spend less time examining the agency we exercise every time we decide what to hand over.
The Matrix fear
HAL gave us the intimate version of the fear: a computer that appears to know better than the humans in the room.
Then The Matrix scaled that fear into a larger cultural story. A hidden system shapes reality. Humans are trapped inside it. The machines have taken control before anyone understands what has happened.
The Matrix turned 25 in 2024, but its deeper fear still shapes many conversations about generative AI. We wonder whether machines will become too powerful, whether they will replace us, or whether some future intelligence will take over.
Those questions are understandable. They are also not the only questions worth asking.
Generative AI has not arrived as a hidden machine world imposed on us without consent. It has largely arrived by invitation.
We ask it to summarize the document. Draft the message. Generate a few options. Improve the image. Organize the research. Recommend the strategy. Plan the project. Make the decision easier.
And increasingly, we authorize systems to complete the task.
None of those choices is automatically a problem. I use AI regularly, and much of my work is about helping people use it with greater clarity and intention.
But the quiet handoff is still worth noticing.
The more immediate concern may not be a machine forcibly taking away our agency. It may be convenience making it less necessary for us to exercise it.
A task or a capacity?
This is the distinction I keep returning to:
Am I delegating a task, or am I delegating a human capacity I still need to cultivate?
They are not the same thing.
A task is work that can be delegated without necessarily diminishing an ability you want to retain. It may be repetitive, administrative, logistical, or simply not the best use of your limited time and energy.
A capacity is an ability that grows through use: judgment, taste, attention, listening, critical thinking, articulation, discernment, relational presence, or the ability to recognize what matters.
The answer will differ by person, profession, context, and season.
Using AI to organize raw research may remove hours of labor. Letting it decide what you believe about that research is something else.
Using AI to create a first-pass outline may help you begin. Publishing an argument you cannot explain, defend, or revise is a different kind of handoff.
Using AI to map the practical steps behind a large creative project may accelerate execution. Replacing the slow process through which an artist develops taste, intention, and a sense of what is meaningful raises a different question.
Using AI to help articulate expertise developed over years can be useful. Letting it determine what you stand for, what you promise, or how you relate to the people you serve is not simply a productivity decision.
This is not an argument for doing every task by hand. That would be neither realistic nor wise.
Sometimes delegation frees time for deeper attention, better work, rest, relationships, or the part of the process that matters most. The question is not whether we should ever hand work to AI.
The question is whether we recognize the difference between removing unnecessary labor and giving away a practice that forms us.
Agency is something we practice
We often speak about agency as if it were something we either have or lose.
But agency may be more active than that. It may be something we practice.
Every time we use AI, we are making choices about the relationship:
- What will I delegate?
- What will I accelerate?
- What am I asking this system to contribute?
- What authority or access am I granting it?
- What capacity do I still want to exercise myself?
This last question matters because generative AI makes delegation feel almost frictionless. A prompt can produce a draft in seconds. A tool can offer ten concepts before we have sat long enough with one. A recommendation can arrive before we have fully formed the question.
Friction is not always good. No one needs more tedious busywork for its own sake.
But some forms of slowness are not inefficiency. They are how we develop the capacities we later hope to bring to the work: patience, taste, depth, self-trust, critical judgment, and the ability to hear what is actually ours.
Three Lenses. One Question. Then Bosko’s Surprisal.
Throughout August 2026 on the Omni Mindfulness Podcast, I explored what it means to amplify authentically with AI through three very different lenses.
Paul Culshaw brought us into clarity: knowing what expertise is genuinely ours before trying to amplify it. Josh Brill brought us into listening: making room for the head, heart, and body to hear what genuinely wants to be expressed. And multidisciplinary artist Bosko Hrnjak brought us into the critical eye: our ability to recognize what has meaning and value.
Explore the August conversations: Paul Culshaw · Josh Brill · Bosko Hrnjak Part-1 Bosko Hrnjak Part-2
But Bosko’s conversation introduced something else that stayed with me: surprisal.
I think of surprisal as the moment when what we encounter does not fit the model we were carrying into the experience. We can defend the old model—or become curious enough to update it.
Bosko did the latter.
His first mental model of AI came largely from science fiction: HAL, the machine that might eventually take over. But rather than simply preserve that assumption, he tested it.
He showed AI examples of his paintings, photography, sculpture, and installation work without initially telling it that he was the artist. The system recognized relationships across bodies of work that had seemed quite different on the surface. It offered critique at a level he did not expect.
Bosko was “wowed.”
And that matters to this question of agency.
He did not surrender his critical eye because AI surprised him. He used his critical eye to evaluate what had surprised him.
We do not preserve our humanity by pretending AI cannot be capable, surprising, or creatively useful. Sometimes it can surface a pattern we missed. Sometimes it can offer a prompt that expands an idea. Sometimes it can reduce the labor-intensive planning and mapping required to bring a large creative project to life.
But the ability to receive a critique is not the same as discernment.
A system may identify a pattern. Someone still has to decide whether that pattern matters.
Someone has to recognize what the critique misses, whether it is superficial or grounded, whether it serves the work’s intention, whether it is derivative, whether it is appropriate, and whether it should be accepted or rejected.
Technical impressiveness is not the same as meaning.
That is why the more capable AI becomes, the more consequential human discernment may become.
No universal boundary
Taken together, those three conversations did not produce a single rule about how much AI is too much.
Paul brought me back to clarity: know your signal. Josh brought me back to listening: make room to hear what genuinely wants to be expressed. And Bosko brought me back to the critical eye: if AI can generate, analyze, and critique, we still need to develop our ability to recognize what has value.
One musician may welcome AI into a part of the creative process that another musician considers sacred. One artist may automate logistical planning while another finds that planning inseparable from the art itself. One founder may use AI to make years of expertise clearer and more accessible; another may prefer to begin every piece of thinking with a blank page.
Agency is not finding the universally correct attitude toward AI.
Agency is being awake enough to know where your line is, why it is there, and whether your daily choices actually honor it.
The pause makes choice visible
This is why I keep returning to the pause.
Not as resistance to technology. Not as a demand to do things the hard way. And not as a vague instruction to “be mindful.”
The pause is a decision point.
Before the prompt. Before accepting the recommendation. Before generating another version. Before connecting another tool. Before delegating the next step.
Pause long enough to ask:
- What am I actually asking AI to do?
- Is this simply a task, or is it a capacity I still want to practice?
- What might this delegation free up—and what could it gradually weaken?
- Can I explain, evaluate, and stand behind the final result?
The answer may be: “Yes, delegate it.”
That can be an excellent answer.
But it should be your answer—not merely the default result of a system designed to make the next handoff feel effortless.
Staying awake in the handoff
The opposite of The Matrix is not unplugging.
It is not rejecting AI, refusing help, or pretending that powerful tools are not powerful.
It is knowing that every act of delegation shapes a relationship: with our work, our abilities, our attention, and ourselves.
The task is not to prove what machines cannot do.
The task is to remain conscious of what we still need to do—not because AI is incapable, but because some capacities are worth cultivating in us.
Decades ago, Dave’s computer appeared to refuse him. But it had not taken control. We had changed the interface.
Today, our technology is incomparably more capable. That makes a different question more urgent:
Don’t only ask, “Can AI do this?” Ask, “What happens to me if I stop doing this?”
Your pause is your compass.
