Philosopher Nick Bostrom used the scenario to illustrate that a superintelligent system optimizing a poorly specified goal can cause catastrophic harm without ever being malicious. It's become shorthand in AI safety discussions for the broader risk of misaligned optimization, not a literal prediction about paperclips.