Over the last several years, working with people across dozens of countries, I have begun to see the same signal surface again and again: technology reshaping how people relate to themselves, to their environment, to their communities, and to each other, built by people who never sat down and worked through what that reshaping would cost. The pattern was difficult to recognize at first, but by now, undeniable.
It seems clear that in the coming years, AI will keep reaching further into how people work, love, raise their kids, and understand the realities of the world around them. But it isn’t clear whether the people this technology touches get a hand in shaping that change, or just inherit it once it’s already done.
In some ways, what’s to come will be uncharted territory, but in others it’s a continuation of a pattern we already know. As recently as the last 20 years, billions of people have experienced the massive disruption of social media, a product that reshaped their attention, their kids’ sense of themselves, and how the public manages its disagreements. All of this happened before society had worked through what that would cost. Nobody sat down and decided social media should be allowed to do that. It just happened, because shipping came first and thinking about the consequences came, if it came at all, a distant second.
I am convinced AI can go differently, but only if we treat that sequence, build first, think later, as the mistake it was rather than the inevitable cost of progress. Even more than what AI can produce, what matters is whether the people it touches had a hand in deciding what it would touch in the first place.
As real as the gains from AI may be, the more meaningful outcome will come from everyone having AI that’s built both with them and for them: technology that helps people do meaningful work, raise their kids, keep their own environment intact, and stay in real relationship with each other, without quietly rewriting the terms of human identity and relationships along the way.
My vision is for AI to be built with the people who use it, not just for them. That means treating what a parent notices about her kid’s AI use, what a worker notices about her trade, and what a community notices about how neighbors treat each other as evidence that belongs in the room where the system gets designed, not as feedback collected once the direction is already fixed. And it means that when the system does cause harm, there’s a clear name attached to whoever was supposed to think that harm through and didn’t.
This is distinct from an industry that builds first and negotiates later when a court case forces the question, and treats the backlash as a safety add-on instead of the risk assessment that should have happened before launch. I believe people living inside a change understand its stakes better than anyone deciding at a distance from them, and that this will only matter more as AI reaches further into daily life.
At its core, this is a question of disruption and consent, and it will only become more important from here.
The question is not whether disruption occurs. Every transformative technology disrupts existing relationships. The question is whether people consent to the new terms of those relationships before they become irreversible. The intersection of disruption and consent may become one of the defining governance challenges of the AI era.
If the current pattern holds, AI will keep redefining these relationships faster than anyone can decide if they are okay with that change, and faster than anyone who owns and controls it is required to stop and ask what it will cost. What’s actually valuable is building differently from the start, with the people a system will touch, not around them, and with someone accountable for having done that work.
I believe people affected by a technology should help define its reasonable boundaries before those boundaries get set for them. That said, reaching real agreement with billions of people raises its own hard questions, and tracing a specific harm back to a specific decision inside a company that doesn’t want it traced is not simple either. Still, I believe building something worth having requires bringing people along as a shared responsibility, not just informing them individually, later, of where AI has taken them.
The rest of this decade seems likely to be the decisive period for whether AI gets built with the people it changes, or simply arrives, unaccountable, the way so much of the last one did.
I believe strongly that AI should be built this way, as a matter of standard practice, not exception. My work is dedicated to building the community and infrastructure that will carry what people are actually experiencing with AI into the rooms where these systems get designed, before launch, not after. That’s the work I’m committed to.
– Angeline