CONCEPT
Human-Machine Interpretive Asymmetry
The structural one-sidedness of human-machine interaction: the human brings rich social intelligence to the encounter while the machine responds procedurally — an asymmetry that deepens rather than closes as AI becomes more sophisticated.
Human-machine interpretive asymmetry is Suchman's term for the structural one-sidedness of interactions between people and computational systems. Humans bring to any interaction the full apparatus of social intelligence: they interpret behavior as meaningful, attribute intention and understanding, read context, and assign significance. Machines respond to inputs according to their programming or their training, without any understanding of the user's situation, intentions, or state. The interaction looks like a conversation. It is not one. It is a human doing all the interpretive work and a machine producing outputs sophisticated enough to sustain the projection. The asymmetry was already diagnostic in Suchman's 1987 photocopier studies and becomes more consequential as the machine's outputs become more plausible.
In The You On AI Field Guide
The concept emerged from Suchman's close study of users interacting with the Xerox photocopier help system at PARC. Users consistently read the system's displays as communicative acts — as if the machine were telling them something, asking them