Ambiently interactive devices hold as many functions and possibilities as spaces, objects, and software. Any combination becomes fodder for speculation. With contemporary applications of ambiently interactive devices in the physical and health space, the knowledge and education domain is sure to follow suit. Imagine something along these lines: Several sheets of film to be adhered to the wall, beveled slightly and matte to blend. Inside and invisible are microcontrollers, an antenna, and the software to power an educational platform for language and rote memorization. Infrared strobes and cameras designed to hone in on retinal reflections of the infrared light are barely perceptible, minuscule gray dots against the translucent material which tell the object when eyes are directed towards it. The objects connect automatically, your private local network indicating that a new unique identifier is activated within its physical perimeter, and you verbally confirm its access rights. You are asked what language you are learning, how long you've been learning it (to which you give a vague answer), and for what purpose. Over the next two weeks, the system shows and repeats both written and verbal vocabulary from the language in question. It tracks how long it takes you to respond, predicting the likelihood of lag in response time being a result of exogenous factors, or a lack of word internalization. Your house becomes a set of flashcards, tracking your linguistic progress and guiding your education appropriately. Their appearance is appropriate, unobtrusive, and bordering on the invisible. Connected to your broader personal network, they don't trigger when the system infers particular states of mind, when company is over, or in an inappropriate context. The device itself merely interprets the signals of the broader system: the house's API indicating that its resident is Busy, Unavailable, or Free to Chat. In turn, it offers an additional sensory node, to be leverged by other devices and services, and increasing the personal network's sensory resolution on the whole. With each new node, the network becomes stronger. Long term personal benefits and reducing barriers to knowledge systems is just one aspect of ambient interaction. Another core component is the nuancing of traditionally direct interaction or decision making points, and the broader uses of these now much richer intersections of person and system. Consider a scenario: A galvanic skin sensor, sewn into the band of an undergarment, is one of many physical sensors integrated into your wardrobe. These sensors have a weak output of a half meter at best, deriving both power and encryption from the phone in your pocket. Like the language learning systems, these sensors are tied into individual, physically aware networks and devices, and provide the hub with appropriate contextual information through these ties and associations. In this case, the sensors are monitoring you while you are shown a new apartment. Your phone knows where it is, cross referencing your calendar appointment with the current wifi triangulation references through Skyhook, and from the sensors against your skin it also knows that you are agitated. Most mobile operating systems come with behavioural models and libraries at this point, with simply accessed APIs for applications and third party processes to apply in broadening the operating system's functionality. One such application is the meAlert. The meAlert is a simple application that runs as a background process in the phone's operating system, and which causes the phone to vibrate in warning and eventually trigger a phone call when user decision making capacity is at a dangerous low. It is a warning to self of self, with the explicit goal of forcing the user to reassess a situation and seek outside advice. Triggering is contextual, based around known locations, physical inputs from the user's worn sensors, and connected services like calendar or online profiles.
The meAlert service is a great free tool for users seeking a personal safeguard, like a preemptive second opinion. Of course, this service comes at a cost. While the behavioral models against which the data inputs are compared is open within the operating system, as we see the OpenCV libraries might be today: these models are effectively only one input in the broader system. A service is elsewhere in place, providing the context and purpose to route physical sensors through the models, deriving meaning from their output. Through the running of this service, meAlert gains a much deeper insight into consumer behaviour and guidance through the path to purchase, and the user helps to naturally grow the effectiveness of future marketing campaigns. This is the danger of ambiently interactive systems. They inherently depend on complex and interconnected networks of data, inputs, outputs, services, and networks. Just with current systems of user profile sharing and tracking, advertising and marketing will become more focused than it has ever been. The lure of data prostitution will be too great. Where we currently trade screen space and tracking cookie for free web services, the future will be access to our homes, environment, and person. But as in the first example, we may well exercise more explicit control over the information and privacy of our homes than ever before: it becomes a matter of culture, behaviour, and design. The role of in defining this new genre of product is that of designing systems and behaviour. While the engineering challenges will quickly be met, some of the behavioural challenges are still far off. The biggest challenge to uptake is the lack of a salient cognitive model for these types of networked object ecosystems. The earlier example of the galvanic skin response sensor on one’s underwear is well and good, provided one understands the value of such an object. Such understanding being lacking, it becomes so much extra copper and annoyance for airport security. Interaction designers are tasked with the role of searching out a mental model for emergent behaviour in disparate but functionally associate objects, and voraciously exploring those spaces. This challenge can be met by several routes: from a business perspective, building systems to make these objects and their use so innocuous as to be common place would become the primary goal. Like the adoption of smart phones, this can permit the creation and propagation of proprietary and shared networks, allowing for the aggregation of situational information across a series of objects and environments within the control of single object ecosystems. This scenario would then see multiple independent networks cropping up, against which one might compare the contemporary platform battles in the tablet or mobile markets. This model allows for the introduction, life, and death of products within a consumer market; a natural cycle of competitive product development. From a product design side, aggressively fusing the design purview of interaction design, architecture, industrial design, software engineering, manufacturing engineering, and electrical engineering (to name a few) under single teams is the first step in making these kinds of networks possible. We are already seeing these kinds of multidisciplinary teams cropping up. Studios like BERG in London, UrbanScale in New York, or Normative Design in Toronto all orient themselves around strong, interdisciplinary teams who speak a common language of design, development, and creation. Fostering shared understandings of the social, psychological, technical, and physical components of these objects and spaces permit a more systematic and creative approach to how these objects might be realized. If current ‘open data’ or open API trends propogate, then a more flexible system becomes feasible. Here, we would see the development of a system which might place identity as the protected value versus that of associate data – permitting personal systems to query, read, and write the physical space without direct fear of privacy violation. This emphasizes the contextual value of a system's data as something that is shared, in the same way that my scuffing a banister or pressing too hard on a pencil affects the history and character of the surface I came in contact with. This data becomes open, accessible, and part of whole environment, but my association with it is anonymized to all but myself.
[Generations of OpenWRT / photo: Security4all] The major benefit to this final option is that it allows for a holistic image of the physical and digital space from the perspective of individual devices and services. Polling and managing of these myriad inputs could be managed by a central hub, a 'cloud' service, or by the device itself. The actual structure of these networks would likely be ad-hoc, self-healing, and valuing flexibility over robustness. Devices and services create an informational history for a place, person, or context. Ones interactions with an environment, and the reactions it has to the individual, become several orders of magnitude more nuanced, allowing for both simplified interactions (though more complex models), and emergent complexity in the types of outputs that become available. We are at the brink. Our phones hold more power than highly capable desktop computers of a few years past, our services and identities are diffused across countless banks of servers, and the sensors already rife throughout our physical environment are slowly becoming aware of their next door neighbors. As the market moves and the consumer reacts, rhetoric and advertising will bring the slow recognition that a sensor in one’s underwear means the difference between 720 and 1080 in their high definition life experience. But not being there yet, there's still time to shape things. To define protocols and standards which protect and enhance personal privacy and freedom. To design and distribute devices and services which allow these borders to be explored. To research and observe the existing software and device ecosystem, looking for those shifts in behaviour or the change in vernacular which might mean an opening door. We are ambiently interactive systems in the end, and this is very much an exercise is figuring out how to build ourselves. |
|||
