| Okay, my mistake in using the term "always-on." That term is generally reserved for broadband network connections. But consider the standard use case for undergarments. It will continuously collect data any time it's worn. That could turn out to be a lot of data, even in a store-and-forward data collection strategy. And really, bras are most commonly worn in social situations, so most data will be collected while the subject engages in social interactions. I point this out as a possible "lie detector" because that's my immediate association with respect to electrodermal and EKG sensors used in combination. Add in respiration (a small leap since this will be strapped to your rib cage) and you would have the complete recipe for a lie detector. That part doesn't seem to be included, although my free association still stands. I'm hypothesizing that Microsoft might be taking aim at a strategy to crowdsource ambient lie detector data, because that sounds like a really interesting data set, with lots of opportunities for exploitation. What motivation could there possibly be, beyond "pure evil", as a drive to study this sort of thing? Perhaps a sense of adventure and bold curiosity for the future that lies ahead? I dunno, but it's already well established that Microsoft is a known collaborator with the DoD and most alphabet agencies. Maybe it's a sureptitious fishing expedition? Why women only? Because a complete lie detector session involves respiration. So, if it's a sureptitious data collection program, you start with the question: how do we convince people to adorn themselves with cumbersome blood pressure cuff, EKG, electrodermal and respiration sensors that we usually wrap around their chest? Who would ever wear anything tightly strapped around their chest? Wait! Women do that all the time, whenever they wear a bra! Good work, Johnson! Here's a two million dollar budget! In six months, be ready to show me something that will convince people to willingly share respiration, heart rate, and skin conductivity via the web. Then we'll restrospectivly analyze the raw data set, blindly without context, and try to pick out the liars, and develop a common statistical model of deceptive behavior, ironically collected using deceptive tactics. |