I was watching the UN session on AI when I felt something familiar and unsettling.
A feeling I associate with watching a particular kind of speech, one that is technically accurate, emotionally amplified, and structurally designed to close the door on the question it is pretending to open.
Two of the most powerful figures in AI were in that room. They spoke about danger, they mostly do and sometimes are not truthful on several occasions. They spoke about the need for governance, for international cooperation, for urgency, implied, without quite saying it, that they were the responsible ones. The good actors. The people you could trust to manage what they had built. They were doing a pitch.
I have heard that rhetorical structure before. Not always from tech leaders. Sometimes from politicians. Sometimes from the kind of personality that history tends to look back on with a particular kind of dread. The alarm is real enough to feel credible. The emotion is calibrated to bypass the part of you that asks, who benefits from me feeling this way right now?
I caught myself feeling it, the pull of it. That low-level anxiety that something enormous is happening and the only people who understand it are the ones already holding it.
And then I stopped. Because I recognised what I was being asked to do.
I was being asked to watch. To scream, spread the word, to contaminate others with fear.
• • •
Here is what the room contained. World leaders, or their representatives. Two AI company founders who answer primarily to their investors. Researchers who have been raising these concerns for years, mostly unheard until this moment. And an implicit question hanging over all of it, who gets to govern this technology, and on what terms?
What the room did not contain, and what nobody in the room named, is the only force large enough to actually answer that question.
Us. The general public.
Not as a slogan or as an abstraction. As a specific, measurable, and almost entirely unmobilised source of power.
Google generates roughly 70 to 80 percent of Alphabet’s revenue through advertising. Meta generates over 95 percent of its revenue the same way. Together they control more than half of the global online advertising market. The entire economic foundation of these platforms, the capital that funds the AI development being discussed at the UN, comes from one source.
Our attention.
Not our data, exactly, though that too. Our attention. The minutes and hours we spend inside systems that convert our presence into targeting signals and sell those signals to the highest bidder. Every scroll. Every pause. Every click that tells the algorithm something it did not know before about what moves us.
That is not a metaphor. That is the business model. And the business model depends entirely on our continued, largely voluntary participation.
So I ask you why they offer free tools?
• • •
The most consequential governing decision about AI is not being made in Geneva or New York. It is being made every morning by a few billion people deciding whether to open an app.
I want to be careful, because this argument can slip into something I do not mean.
I am not saying individual choices are enough. I am not saying that if we all deleted our accounts tomorrow, the problem would be solved. The systems are too embedded, the economic dependencies too real, and the cost of disengagement is not equally distributed. For many people, these platforms are not optional. They are income, visibility, connection, livelihood, their lives, as we got use to.
What I am saying is different. I am saying that the framing of the UN session, powerful actors governing a technology on behalf of a passive public, is not just incomplete. It is structurally convenient for the people in the room.
Because if the public is passive, then governance happens to us. Rules are written by the people who already have power, enforced through mechanisms those people designed, and adjusted when those people decide adjustment is necessary. The rest of us receive the outcome.
But if the public is the source of the economic power that makes all of this possible, then the question changes. Governance is not something done to us. It is something we are already participating in, whether we know it or not, through every decision about where we place our attention and what we refuse to fund with it.
The problem is not that we lack power. The problem is that we have been very carefully taught not to see it.
• • •
I recognise the rhetorical move I described at the opening because I have used versions of it myself.
Not at a UN session, but in product meetings. In roadmap presentations. In the language we used when we were deploying systems faster than we could fully understand their effects, and we needed the people who might ask hard questions to feel that the situation was too complex and too urgent for hard questions.
Alarm, when used correctly, is not a warning. It is a closure mechanism. It says, the stakes are too high for the normal checks, trust the people who understand this, we do not have time for the process you are about to suggest.
I have sat in rooms where that move was made deliberately. I have sat in rooms where the people making it believed every word and still did not notice what the structure of their argument was doing. The effect is the same either way. The door closes. The critical question does not get asked. The decision gets made by the people already holding it.
What I felt watching the UN session was the same door closing, at a much larger scale, with much higher stakes.
And I noticed, because I have learned to notice it, that the speech was designed to make me feel emotion before thought. Urgency before scrutiny. The appropriate response to danger before the question of who defined the danger and why.
That is not the grammar of a genuine warning. That is the grammar of consolidation.
• • •
When someone tells you the situation is too dangerous for you to think slowly about it, that is usually the moment that most requires you to think slowly.
What would a different session have looked like?
Not a session that excluded the AI companies, their technical knowledge is real and relevant, not a session without urgency, because some of the risks being named are genuine. But a session that started from a different premise.
One that treated the public not as the population to be protected, but as the constituency with the most direct stake and the most latent power. One that asked not just what rules should govern AI, but who gets to write those rules, who enforces them, who can challenge them, and what happens when the people enforcing them are also the people who benefit most from the current arrangement.
One that named, openly, that the economic foundation of the most powerful AI systems on the planet is our attention, and that this creates a leverage point that no government delegation and no corporate governance framework has yet seriously engaged with.
One that asked what it would mean for us to govern, not just to be governed.
That session did not happen. I do not think it was an accident that it did not happen. The people with the most to lose from that conversation were the people with the most influence over what the session contained.
• • •
I want to end with something I cannot fully resolve.
The argument I am making, that we are the source of the power being debated over our heads, only holds if we are willing to act like it. And that requires something that the current design of these systems makes very difficult: the capacity to think slowly, in the face of constant pressure to react quickly.
Digital fatigue is not a personal failing. It is a structural outcome. The systems we are inside were built to keep us in a state of low-grade, continuous alertness. Cortisol and attention residue and the splintering of intention. The inability to hold a thought long enough to follow it somewhere the algorithm did not already plan for you to go.
That fragmentation is not a side effect. It is a feature. A fragmented public cannot coordinate. A public that cannot coordinate cannot govern. A public that cannot govern gets governed.
And so the most radical thing I can suggest is also the most ordinary. Think slowly. Finish the thought before you share it. Notice the rhetorical structure of the alarm before you feel the alarm. Ask who benefits from you responding with your nervous system rather than your judgment.
Not because individual choices will solve structural problems. They will not, and anyone who tells you otherwise is selling you a responsibility you were not designed to carry alone.
But because the power that nobody named in that UN room starts with the capacity to see it. And you cannot see it if you are too fragmented, too fatigued, too conditioned to react, to look.
The most powerful people in that room were not in the room.
They were watching it. On their phones. Between notifications. Half-present.
That is the thing I cannot stop thinking about.
If we are the source of the power being governed over our heads, what would it actually mean to stop watching and start deciding?
About the Author
Diamantino Almeida is a tech leader, coach, and writer reshaping how we think about leadership in a burnout-driven world. With over 20 years at the intersection of engineering, DevOps, and team culture, he helps humans lead consciously from the inside out. When he’s not challenging outdated norms, he’s plotting how to make work more human, one verb at a time.