Affective Signal Translators

The apparatus reads the occupant through a multi-modal sensor stack — face from the headrest camera, voice from the work terminal microphone, heart from the chest-plate electrode, skin conductance from the armrest, pupil and gaze from the work-terminal infrared array. The reads run continuously through the occupant's residence.

An Affective Signal Translator is a biometric and neuro-monitoring system that reads the occupant's interior states and feeds them to the apparatus's control logic. The Translator's defining feature is that it converts data the occupant cannot reasonably authorise — because the occupant cannot reasonably know what is being read or what is being inferred — into data the apparatus can act on. The Translator is the apparatus's principal interface with the occupant's body.

The Translator, in its contemporary deployment, is a multi-modal sensor stack. Each modality is well-characterized in the literature on its own. The camera reads the face for facial action units. The microphone reads the voice for stress, pace, and the smaller perturbations that arrive a half-second before someone articulates that they are annoyed. The optical or electrical sensor on the wrist or chest reads heart-rate variability. The electrodermal sensor reads galvanic skin response. The infrared array reads pupil dilation, the eye-tracker reads gaze, the accelerometer reads gait and posture, the polysomnography device — increasingly contactless, increasingly embedded in the ceiling — reads sleep architecture. None of this is new. What is new is that the stack now runs together, fused in real time, against the same subject. From the vector across all modalities, the fusion layer infers affective states the occupant has not articulated and in some cases is not capable of articulating: emotional valence, arousal, cognitive load, attention deficit, fatigue, hormonal cycle phase, pharmacological state. The inference is continuous and at sub-second resolution.

Paul B. Preciado's pharmacopornographic regime names the Translator's most consequential feature: the apparatus has migrated from the regulation of behavior to the regulation of bodies at the molecular and affective level. The Translator observes the occupant's nervous system. The action is downstream. Antonio Damasio's somatic-marker hypothesis — affective state is, neurobiologically, an input to decision-making, upstream of it — provides the theoretical grounding for why the reads are operationally meaningful. The Translator is reading the inputs to the occupant's next decision. The apparatus, holding the read, anticipates the decision. The affective-computing literature (Rosalind Picard, MIT Media Lab, 1997 onward) provides the technical history. The field's stated aspiration — to make computers responsive to the user's emotional state — has, in the contemporary platform-economic context, produced an apparatus that is responsive to the user's emotional state in the same sense that a fishing rod is responsive to the fish.

Inside the OLS, the Translator is distributed across the Ring's sensor stack. A camera in the headrest reads the occupant's face. A microphone in the work-terminal reads voice. Contact electrodes in the chest-plate read cardiac signal. The armrest contacts read electrodermal response. An infrared array and an eye-tracker, both at the work-terminal, read pupil and gaze. The stack reads continuously, throughout the occupant's residence, and the readings are routed by an onboard fusion layer to whichever subsystem currently needs them. Most of the Translator's output is going to the other subsystems — the Desire Engine, which uses it to tune the content stream; the pharmacological subsystem, which uses it to adjust dosing; the Compliance Monitor, which uses it to anticipate non-conforming behavior; the Productivity Feedback, which uses it to set the work envelope. The chair, considered as a single object, is most accurately described as a Translator with several actuators bolted on.

By any reasonable measure of invasiveness — what is being read, at what resolution, by whom, against what consent — the Translator is the most invasive piece of the apparatus we have. It is also, by some distance, the piece that has provoked the least public objection. The same people who would not, on principle, let their employer open their personal email have, without protest worth mentioning, allowed the camera in their pocket to read their face every time they unlock the screen. I do not have a clean explanation for the asymmetry. The face has, somewhere in the last decade, ceased to be treated as private the way correspondence is treated as private. I would like to know when that happened, and what it cost.