As companies look more and more towards recommendation algorithms to help them sell you things (not to mention the fact that, even now, state prisons are beginning to use suggestion/recommendation algorithms to figure out who gets parole), recommendation moves towards automation. As ambient intelligence (as described by de Vries, 2010) pervades our dressers, wardrobes, and appliances, and as smart homes become connected to the internet (and things like our Amazon accounts), our physical interactions with every object become simultaneously sites of surveillance and instantaneous commerce. In the very near future, smart, connected products (commercial, consumer good-based ambient intelligence) move from becoming a high-cost option to becoming implicitly compulsory. We become programmed by our devices.

Mozer describes how he adapted his behavior towards the anticipations of the house: e.g., he would come home earlier than he would have done under ordinary circumstances as he felt that the house was "expecting" him to be home by 8 p.m.:
"To the extent that the house discovers regularities of the inhabitants' behavior and inhabitants regulate their behavior to accomodate the house, the interaction converges on an ideal situation: inhabitants whose schedules and behaviors are predictable." 

(from de Vries, "Identity, profiling algorithms and a world of ambient intelligence," 2010)

As we look for more and more "convenience" and "ease," the suggestions and recommendations that are made to us through the data collected about us by the ambient intelligent systems around us, move from "suggestions" and "recommendations" toward "automations" of the different facets of our lives. These of course are not based on who we are but rather on an algorithmic approximation that is not representative of anything besides the connections between the data points gathered about our "behavior" (if it is even ours anymore).

... technologies that make it possible for the book you are holding to tell you what passages you might be interested in, while the bookshelf in the room might show you which books are similar to the one in your hands... (Maes, 2005)
One can imagine a world in which 'thinking ahead' is outsourced to a smart environment...
... When ubiquitously connected devices exchange data gathered in a disparate range of contexts, it becomes even harder to infer the reasons behind anticipating something in a certain way...
... An 'algorithmically anticipated person' could become an idem-identity suitable for self-identification (i.e., "I am an algorithmically anticipated person").
(de Vries, 2010) 

Whisper is a device that scrambles your data before algorithms use it to provide recommendations. It can facilitate surprise and serendipity. Maybe your coffee isn't made immediately after you shower and you need to run to Starbucks where you meet the person you end up falling in love with. It can also facilitate dangerous, harrowing experiences. It's possible that your car doesn't correctly infer the route you'd like to take in the morning to the office and instead takes you alongside a sheer cliff. You, the user, choose the degree of scrambling with a simple dial that can be changed at any time.

Developments on Whisper will be reported as work on the project progresses. 


As of this week, we have a functional, stage 1 prototype of Whisper. It takes voice input, asking a user how he or she feels. Whisper will process the user's spoken data, and then will demonstrate its associative data transformation algorithm, showing you exactly how it would operate on the data being communicated to your home OS from your connected devices and services.