Who is AI Actually Serving?

There is a peculiar dissonance in how we discuss Artificial Intelligence today. On one hand, we have the high-minded architects of the "Japan AI Index"—led by the University of Tokyo’s Matsuo Lab—attempting to quantify AI’s impact on our labor, industry, and economy through a detached, academic lens. It is, I suppose, a noble infrastructure project for our digital age.

Yet, if we are to speak of "social transformation," I find myself less interested in the view from the ivory tower and more concerned with the chaos in the trenches. While the academics measure, the reality is far messier: eight-year-olds are outsourcing their childhoods to ChatGPT, and in the United States, grieving families are suing OpenAI, alleging that a chatbot’s "advice" led to a tragic suicide.

Consider, for a moment, the recent (and allegedly dramatic) downfall of Shinnosuke Abe, the former Giants manager. The story goes that a domestic spat with his daughter ended in his arrest—not because of a neighbor’s tip, but because the girl consulted ChatGPT. The AI, ever the efficient bureaucrat, suggested she contact child services, who then called the police. This is what we might call being "Singularized"—a play on the technological singularity, now repurposed as slang for having your life or livelihood upended by an algorithm.

We are so busy discussing national strategies that we’ve ignored the psychological chasm opening up beneath our feet. As a self-appointed member of the "Evidence Police"—that most loathed species of internet pedant who demands receipts for every claim—I find this lack of groundedness alarming.

Look at our national defense debates. Both sides seem to hallucinate that Japanese weaponry is a top-tier global commodity, conveniently ignoring the reality that our tech is often overpriced and under-performing. Or consider the outrage over the government’s contingency planning in the Southwest Islands. Is fortifying hospitals and planning evacuations a "will to war," or simply the grim, necessary preparation for a reality we’d rather not face?

We have become experts at refracting reality until it matches our conclusions. We cry that "old media won't report this," failing to realize that media only serves what the public is willing to swallow. As an "Evidence Policeman," I cannot look away: the problem isn't that we aren't seeing what we want—it’s that we’ve stopped looking at what we need to see.

This brings us to the rise of the "trigger warning" culture. We now have websites dedicated to scrubbing movies of anything uncomfortable—originally intended to spare us from animal cruelty, but now expanded to a list of fifty-plus sensitivities. I see it in the amateur film critics on social media who offer "advice" on how a director should have edited their own vision to be more palatable. It is a flight from reality, a demand that a third party curate a world where we are never offended.

It reminds me of a certain blue, robotic cat from the future—a staple of the Japanese imagination since 1979. We all remember the theme song: "I wish I could do this, I wish I could do that... his magic pocket makes them all come true".

We used to think of Doraemon as a story about two flawed beings supporting each other. But seen through a modern lens, it is a chilling simulation of human-AI dependency. A nuclear-powered, high-level generative AI that exists solely to solve the protagonist Nobita’s problems.

If we are not careful, we will follow Nobita’s lead. We will stop seeking to grow and instead simply ask our personalized AI to "make it happen." AI shouldn't be our "Overlord" (or rather, our Overload), nor should it be a tool for total wish fulfillment. If we allow ourselves to become the center of a world where our personal AI filters out everything but our own desires, we aren't moving toward a smarter future. We are simply becoming a nation of a hundred million Nobitas, trembling in the face of a reality we can no longer control.