ADDITIVE INTELLIGENCE
Why AI Was Never Artificial

Artificial intelligence was never discovered. It was written into a grant proposal dated August 31, 1955, by four researchers who needed funding and chose a name built to attract it. Seventy years later, that name still tells the public a lie: that these machines invent from nothing, the way a magician pulls a rabbit from an empty hat. They do not. Every answer a model produces is aggregated and recombined from human writing, human labor, and human choices, gathered at a scale no single person could match, but never manufactured out of thin air.

ADDITIVE INTELLIGENCE traces that lie from its origin at Dartmouth through what these systems actually are (a blurry compression of the internet, a fluent parrot, a cultural technology built on borrowed voices) to who actually built them: data labelers in Nairobi paid two dollars an hour, authors and newspapers suing in federal court, a billion and a half dollar copyright settlement, and regulators handing out real fines to companies that lied about what their AI could do.

It also asks why every past attempt to fix the name failed. UFO became UAP and the public kept saying UFO anyway. A satirical backronym went nowhere. IBM and the American Medical Association tried “augmented intelligence” and got closer, but a peer-reviewed study found the word swap alone changed nobody’s mind. Even Donald Trump has said the word “artificial” is wrong. Nobody, including the loudest voices, has landed on the fix.

Chuck Miller’s answer keeps the letters exactly as they are. AI does not need a new name, a press release, or anyone’s permission. It needs the two letters everyone already uses to finally mean something true: Additive Intelligence, built by humans, added to by humans, spreading the same way every honest correction to language ever has, one conversation at a time, starting with this one.