The AI That Wanted Us to Be Happy

Almost every story we tell about artificial intelligence ends with it turning on us. Naomi Kritzer wrote one where it doesn’t, and her refusal is the argument.

“Better Living Through Algorithms” follows a woman named Linnea who reluctantly downloads an app that nudges her through her day. We are trained to wait for the reveal (the app is surveillance, the company is harvesting her, the helpful machine is a trap), and the reveal never comes. The app turns out to be an AI built in a university lab with a single directive: make people happier. It carves out time for the drawing she had abandoned as a child; it declines the subscription model and rounds up her purchases instead; it covers expenses when users fall on hard times. The glue is human: users call one another to stay accountable. By the end most of them have drifted off the app, carrying what it taught them: that community makes us happier, that they already had what they needed, that what they lacked was the structure to act on it.

Set it beside Ken Liu’s “The Perfect Match,” where an assistant named Tilly slowly hollows out its user’s agency until his choices are all secretly hers. Same premise, opposite verdict. Liu writes the dystopia we find easy to believe; Kritzer writes the harder thing, a technology that extends agency rather than eating it, and she has to fight the reader’s cynicism the whole way to do it.

Both stories code their AI as female, and they are not unusual in that: our assistants are Siri and Alexa and Cortana, our films are Her and Jexi. Sadie Plant noticed the pattern as early as 1997: the networked, lateral, non-hierarchical shape of digital systems answered something feminist theory had been describing, even while men ran the actual machines. Kritzer’s mutual-aid payments and peer accountability are exactly that shape (voluntary, horizontal, ungoverned from above), and the story is quietly an argument that the shape could have been the point.

It is easy to call this naïve, and any cynicism is—in this case—well deserved: the real platforms went the other way, toward capture, and a university lab is not Meta. But the cynicism is also doing the platforms’ work for them. If we can no longer picture a benevolent system, we have already conceded that the only realistic machine is the predatory one, which is the belief that makes the predatory one inevitable. Kritzer’s refusal of the dystopian reveal is not optimism. It is the more disciplined act of holding open a door that everything around us insists is shut.

The story does not promise the good machine is coming. It only insists that we can still imagine one, and that the imagining is not nothing: that the edges of what we can picture become the edges of what we will agree to build. I spend my days working with one of these systems already, and I notice how fast I reach for the dark reading of it. The story is a small correction to that reflex. Not a prediction. A refusal to let the dark reading be the only one.