Post

Trump’s switch from “artificial intelligence” to “super intelligence” creates a terminology problem. In AI research, superintelligence already has a specific meaning: a hypothetical system whose capabilities substantially exceed humans across a very broad range of domains.

“Artificial” in artificial intelligence doesn’t mean fake. It means the intelligence is produced by an artificial, human-made system rather than a biological organism. Artificial satellites aren’t fake satellites. Artificial limbs aren’t fake limbs.

And “superintelligence” isn’t a compliment meaning “really impressive AI.” It’s a capability classification. Calling today’s AI superintelligence doesn’t give it those capabilities any more than renaming a sports car a “hypersonic vehicle” makes it travel at Mach 5.

If everything from a chatbot on your phone to a frontier reasoning model becomes “super intelligence,” what do we call the thing researchers have historically meant by superintelligence? A terminology change isn’t evidence that today’s AI has crossed that technical threshold.