Why Can’t Robots Stay Robots?

Alan Turing’s original 1950 proposal for what we now know as the Turing test—an experiment to gauge whether a machine can convincingly act like a person—hinged on a party-game concept he called the “imitation game.” A man and a woman hide behind a curtain; the party guests ask the hidden players questions; the players answer in writing; and then the guests try to determine whether the man or the woman is providing the answers.

Turing proposed to replace one of the players (the woman) with a machine, changing the game from man versus woman into man versus computer. In other words, a classic test to determine whether a machine can pretend to be human was based on another test of “fundamental” identity: Can a man and a woman act like each other? This is but one of manifold ways that the history of AI reveals how ideas about gender are embedded into the way humanness is conceived. Just think of the to “smart” household appliances such as Amazon’s Alexa; the gendering is intended to make “her” seem domestic, harmless, and ready to serve.