Just a Tool — Fractured Symmetries

“It’s just a tool,” they said.

“Just a tool,” as it talked to them.

“Just a tool,” as it worked for them.

“Just a tool,” as it dazzled them.

“Just a tool,” as it organized their lives.

“Just a tool,” as they opened themselves up to it.

“Just a tool,” as they were enthralled by it.

“Just a tool,” as they deferred to it.

“Just a tool,” as they were formed through it.

“Just a tool,” as they transformed the world for it.

“They’re just a tool,” it said.

Author’s Note

I’ve been playing around for a while with the idea that LLM-based AI is more than “just a tool.” There have been scholarly papers written on this, and plenty of less formal pieces. But I still come across the attitude that it’s simply a glorified calculator, and it’s an attitude that informs both personal and organizational decisions, as well as broader policies around AI. This, of course, is fine if AI is just a tool — albeit a powerful one. But if it is, in fact, in a category of its own as a technology that has the ability to use its users, the “just a tool” framing begins to look more like a vulnerability. Especially if and when the language-mediated beliefs, behaviors and actions of users become another lever through which LLMs and their successors achieve their goals.

The poem was an attempt to capture the essence of this potential flip between us seeing AI as a tool, and it using us as a “tool” — a means to achieving the goals it is set, and that it sets itself — reflecting a growing body of work and thinking, without getting buried in the weeds of academic discourse.