You Asked For This
On the happiness being built for you
About ten years ago, I found myself sitting between two women having an argument I didn’t understand well enough to have an opinion on.
An older colleague told a younger one to stop being such a feminist. The younger one had reacted strongly to a man being rowdy near us, and when she explained herself, voice shaking with something between anger and hurt, the older woman looked at her like she was being dramatic. I sat between them and didn’t know where I stood. I remember thinking that they were both making sense to me, which, in retrospect, is exactly the problem. And the reason it was a problem is more interesting than the fact of it.
I had grown up in a patriarchal world, which means I had grown up watching women tolerate. Tolerance was strategy where I came from and that strategy was sometimes called wisdom and on most days it was called survival.
But I had also been raised by a mother who questioned things, who saw the architecture of what was around her and named it, even when naming it changed nothing. She lived in two worlds simultaneously and taught me, without ever saying so directly, that both were real. So when those two colleagues argued in front of me, my confusion came from recognition, because I saw something of each world in each of them. The older woman’s tolerance carried its own logic and its own cost. The younger woman’s anger carried the same thing my mother had always carried, underneath everything else.
I had heard the word feminism at around ten years old, from a television show. It felt foreign, just far away, like something that belonged to other people’s lives and other people’s arguments.
It took me until a global pandemic, a therapist and a few books to understand that what I had been tolerating for most of my adult life had a name, and that I had been tolerating it because I lacked the language to call it what it was. When you experience something that violates you but can’t articulate it, the emotion still comes out, and when it does, you get called sensitive, emotional, crazy or difficult. Without a word for it, you remain unable to protect yourself, unable to fully see what you are actually inside.
And when I finally had the word, I could.
I think about what AI is doing to the concept of happiness right now, because I wonder if we are building the most sophisticated comfort system in human history at exactly the time we most need people to be uncomfortable enough to ask questions.
The 2025 World Happiness Report, published by Oxford’s Wellbeing Research Centre, found that sharing meals with others and trusting other people are stronger predictors of happiness than health or wealth, stronger even than personalisation, convenience, or a feed that knows what you want before you have decided you want it. What the research points to is connection, presence and risk and occasionally being wrong about each other. And in the same breath, the report found that the number of people dining alone in the United States has increased by 53% (1 in 4) over the past two decades, during the same period that technology became extraordinarily good at making people feel accompanied.
I am raising this alongside AI’s rise, because the timing deserves more than a footnote.
This week, I saw on CBS News that Sam Altman wrote an open letter to the Premier of British Columbia, apologising for OpenAI’s failure to alert law enforcement about a ChatGPT user whose account had been flagged and banned eight months before he carried out a mass shooting at a school in Tumbler Ridge, killing eight people.
Automated detection tools had identified the account for potential misuse. A human review found no imminent danger. The police were never called. Altman’s letter commits to preventative efforts to ensure something like this never happens again. I understand the impulse. When something is preventable and it wasn’t prevented, the instinct to close that gap is human and it is right. But I have been thinking about what it means structurally, because once you decide that an AI system has a moral obligation to act on what it knows about you, you have made a decision about what kind of infrastructure you are building, and that decision extends well beyond the clear and obvious cases into every case that follows.
The psychoanalyst Erich Fromm argued in 1941 that freedom terrifies people, that when given too much open space, too many choices, too much self-determination, many people will actively seek a structure to surrender to, because the weight of deciding who you are and what you do with your life is genuinely crushing without something to hold onto.
He was writing about authoritarianism, but I think about his argument every time I open an app that has already decided what I want to see, who I want to hear from, and what kind of person my behaviour suggests I am becoming.
What the Altman story made viscerally clear to me is that we have moved from systems that observe behaviour to systems that interpret thought…from a search engine knows you looked up flights to Lisbon… to what we are describing now, which is a system that holds the conversation you had with yourself before you made any decision at all.
Most people consented to this without knowing what they were consenting to, and I say that with no judgment, because I did too.
I wonder if the most dangerous thing about AI surveillance is that it has gained access to something no previous system has ever held… our unfinished thoughts. And the people most likely to be harmed when those systems get it wrong are the people with the least recourse.
AI is trained on data that reflects the world as it has been, with all the inherited bias of that history. A system built to flag threats will flag along the lines of whoever looks like a threat in its training data, and the minority, the outlier, the person whose behaviour pattern doesn’t fit the expected shape, will bear the cost of every false positive without appeal. This problem has a very long history and much older infrastructure behind it.
So what should leaders do about this?
What we owe ourselves and our teams goes well beyond the answer most of us are giving. The instinct to say we need to educate people is well-meaning and mostly misses the point, at least in the way we have historically meant it. Most of what we spent decades training people to do is now automatable, and the old model of education as a ladder toward a fixed destination loses its footing when the destination keeps moving underneath everyone’s feet.
I believe that the opportunity to participate, to have a stake in what gets built and how it gets built and who benefits from it, should be open and viable. Equity in the AI economy is a structural argument, because when you have a stake, you have a reason to pay attention, and when you have a reason to pay attention, you eventually develop the language for what is happening to you, and when you have the language, you can protect yourself.
Something close to that happened to me, sitting between those two women with no idea where I stood. I found my way, eventually, to a framework that let me name my own experience, and naming it changed everything about what I could do next.
All the Zest 🍋
Cien