Ambient intelligence envisions a future in which our environment is populated by an infinite number of interoperating, computing-embedded devices of different sizes and capabilities , which are interweaved into "the fabric of everyday life" and are indistinguishable from it . The ultimate goal of all these devices is to serve human needs through the provision of a wide range of physical, digital, and hybrid services  that improve the quality of life by making it easier (smart homes, e-commerce), safer (accident prevention and avoidance, monitoring the location and safety of children), healthier (assisted living, telemedicine), more productive and efficient (teleworking, traffic management, driver assistance), and even more pleasant (social-interaction and entertainment activities).
The first step toward meeting any human need is, of course, identifying it. Ambient intelligence technologies achieve this task through a vast collection of hardware and software modules, which can generically be described as "sensors" , since, in one way or another, they have the ability to collect information that is, implicitly or explicitly, produced by humans. These various pieces of information can then, individually or collectively, be mapped to human actions, states, intentions, and, eventually, needs.
Examples of hardware sensors include simple mechanical or electronic devices, such as microphones, video cameras, distance, movement and pressure sensors, as well as more sophisticated apparatuses such as electroencephalographical devices and neural implants. In general, all these sensors detect physical "output signals" of the human body, whether sounds and gestures, physiological measurements, exerted forces, electrical signals, or brainwaves.
Software sensors complement the hardware ones but do not have a physical embodiment. Their role is to detect the immaterial (i.e., intellectual) products of human activity created through the mediation/support of information and communication technologies. For example, these sensors may monitor Internet-based services such as email, Web portals, chat, and search engines, but also operation systems or typical desktop applications, like word processors, spreadsheets, and computer games.
The information gathered by the various sensors is then propagated to software modules thatassuming the role of miniature brainsstore it, analyze it, change their own internal status and then decide accordingly upon related actions that should be taken, often affecting the state of the user's physical or digital environment. In a future ambient-intelligence-enhanced world, there may be several such decision-making modules of different sizes, capabilities, and sophistication working independently, collectively, or even antagonistically.
Generally, unlike the approach of Big Brother, described in George Orwell's 1984, it is expected that in the emerging ambient intelligence environments, most of the collected information will be distributed among the numerous sensors and applications (i.e., many "small brothers"), never making it to a centralized repository of any kind; also, due mainly to storage constraints, a large part of it will never be permanently stored.
But what if, at some point in the future when practical barriers are alleviated, the option of accumulating each and every piece of information, no matter how trivial or elaborate, related to the whole life of a single human being becomes feasible? Would we then be able to re-create a complete and accurate representation of that personnot only her appearance, actions, interests, and habits but also her personality and way of thinking?
And what if, instead of feeding this model into software applications and services, we presented it to its rightful human owner? In this case, the biggest obstacle that we would probably stumble upon would be that of data rendering, or how to present such massive and diverse, and potentially chaotic, data in a structured, meaningful, and comprehensible way.
For the sake of our hypothesis, let's suppose that we eventually manage to create an appropriate rendering machine, which we entitle the Ambient Mirrorthe contents of which are accurate, noncontroversial, and human-readable.
Then, the first thing that one would probably ask is, "Will the person in question be able to recognize their actual self in this digital reflection?" You see, humans tend not to keep a "high resolution" account of who we are, what we think, or what we do. Instead, we draw a rough sketch of ourselves based on a small, and sometimes fictitious, portion of the available information. Our brain constantly filters millions of details, retaining only a few important facts or events, often creating idealized, mitigated, or aggravated versions of them. Furthermore, over time, much of this retained data is forgotten, corrupted, pushed back, or merged.
If we somehow manage to get a satisfying answer to the aforementioned question, then the next step would be to identify possible "applications" of such a tool and assess their potential usefulness, as well as their impact both to the individual and the to the society as a whole.
On an abstract level, the Ambient Mirror can be described as a tool for recording past personal experiences. Over the course of a few thousand years, humans devised and used several other tools for the same cause. Indisputably, the first one was memory. Then came the spoken language, writing, as well as art, mainly in the form of painting and sculpture. In the modern age, the task is further facilitated through a multitude of electronic devices such as microphones, photographic/video cameras, and computers. Of course, all these "traditional" means will also be basic constituents of the Ambient Mirror, but in a novel, pervasive (and probably miniaturized) manifestation, as well as integrated with additional monitoring technologies and reasoning components that will extend their grasp and complement their abilities. Table 1 illustrates an attempt to sum up the foreseen differences between past practices and the Ambient Mirror.
Based on its unique characteristics, some of the possible personal uses of the Ambient Mirror include:
Extending human memory and awareness. For anything ranging from today's schedule or self-medication plan, to vacation memoirs or critical incidents of one's life. In this case, the mirror works like a recording/ playback device that can present information to its owner both (pro)actively, based on predefined programs ("always remind me about a meeting an hour ahead," "when I meet someone, remind me of her name") and passively, using dynamic requests ("I want to re-experience: (a) X specific minutes of yesterday; (b) all the times that I felt happy during the past month; (c) a collection of my personal thoughts expressed today"). Since the mirror collects data from far more sources than the human senses, it can potentially "remember" more things about the self than this person can even perceive, thus not only increasing the storage capacity of the human brain, but also (retrospectively) enhancing its awareness capabilities.
But if we rely even more on technology for remembering things, would we then deliberately further weaken our own memory? Could we arrive to the point where we use our brain only for short-term storage of data needed for immediate processing and then dump everything else on the Ambient Mirror, in a way that's analogous to RAM and hard disks? And what will happen if the mirror "breaks," for example, it is partially destroyed or is infected by a virus? On the other hand, if the mirror helps us free all the memory space and processing overhead devoted to mundane things, would we then be able to put these resources to better use? Beyond these concerns, another question is, would the capability of accurately replaying past events result in an endless torment of regret about things that we should (or should not) have said and done?
Seeing the unseen. In other words, converting abstract personal information into a tangible form. The mirror intrinsically possesses this ability since most of the collected data is imperceptible to human senses (physiological measures, analysis of written and oral communications). This information can work as real-time feedback for various uses such as stress control, accident prediction and avoidance, or instant health or mental status checkup before an important match, performance, or meeting. A related example is that of biofeedback, in which bodily functions (blood pressure, heart rate, skin temperature) are measured and then visualized (using numbers, graphs, or even games) in order to help a person understand and master her unconscious physiological activities. Furthermore, the mirror has the ability to keep track of, spot, and report phenomena that are usually difficult to self-identify because they span or change over time, such as habits, interpersonal relationships, likes, dislikes, acquirement of (or loss) of skills, degradation of senses, etc.
Self-knowledge and self-improvement. Development of physical skills (much like what athletes and actors do by reviewing videos of their performance to identify errors and correct their technique) and introspection/self-reflection for scientific, educational, and psychological purposes. A basic concern of this type of personal use is that currently some guidance or monitoring by a trained specialist (a trainer or a psychologist, depending on the case) is required in order to ensure the safety of the person in question and the achievement of tangible results. In this case, one may claim that required expert knowledge can also be offered by the Ambient Mirror, either in the form of intelligent agents or by acting as a mediator between the person and remote experts. Another potential risk is that the mirror may work as a self-fulfilling prophecy; the person looking into it may (sub) consciously alter their actions or even character traits in order to "align" themselves to their perceived image.
If, as admitted earlier on, we can really see ourselves in the Ambient Mirror, then the mirror probably would also be a handy tool for allowing other people to "see" us. Currently, this function is served by personal conversations, profiles, CVs, biographies, blogs, photo albums, and video clips. The key difference with the mirror is that it will not just allow you to see somebody else, but to almost experience being somebody else!
Of course, in this case, one would like to be able to decide who has access to which parts of one's mirror. In other words, some kind of "tools" will be needed for selecting specific parts of the mirror's content and releasing them to selected viewers. An ethical question raised here is related to the fact that, as a person's life path (intentionally or accidentally) intersects at some point in time with that of other people, the mirror will also contain data referring to them, compromising their privacy. A possible way to overcome this problem is through a pervasive filtering mechanism, which will be able to anonymize, block, or even erase data recorded by all or selected other Ambient Mirrors. Such a solution may create new problems, since one can foresee that this function can easily be exploited for "stealthing" malevolent actions. Additionally, if someone managed to get control over other people's filtering mechanisms, they could apply a novel form of censorship that can reach to the extent of virtually erasing all evidence of a person's existence.
The related potential applications of the Ambient Mirror include all those situations in which personal information needs to be disclosed, such as interpersonal relations and communication, work applications, and medical filing, among others. Furthermore, the mirror could advance the art of autobiography to a new level, since it will allow the "reader" to almost relive the "writer's" life, experiencing a kind of "living reincarnation."
But what if the mirror is also used as a piece of evidence, like in a courtroom? Would anyone be able to use fragments of their mirror as a valid testimony? Is there a possibility that someone manipulates the content of your (or someone else's) mirror for someone else's favor? Would a third party (the state, the police) be allowed access to anyone's mirror, thus materializing part of the Big Brother scenario?
And, since, in our example, the mirrorup to this pointis considered a personal artifact, whose interests should it serve when a conflict arises? Its owner's or those of the "universal" truth? In other words, would my mirror lie for me? Would it give me away to my boss when I ask for a day off pretending to be sick or to my wife when she asks me where I was last night? And what about all these cases of "white lies" imposed by political correctness, for example, when one asks us if we really liked the food, the dress, the painting, or if we had a good time? Would the mirror become a tool of unprecedented rudeness?
In thenot so distantfuture, it is quite possible that ambient intelligence technologies will provide scientists with all the components and knowledge required to build a device similar to what is described in this article as an Ambient Mirror, reflecting every trace of human existence. Ideally, this mirror will become a new means of altering human self-perception, as well as that of others, eventually leading to autognosis while also fostering mutual respect and understanding. Furthermore, the mirror may even develop into a kind of "synthetic conscience," shepherding people when they go astray. Then again, if the mirror is not working or properly used, it could totally distort our view of the world with unpredictable, but unquestionably catastrophic, consequences. Furthermore, the mirror might prove to be the ultimate surveillance mechanism, which would permanently end the notion of privacy as we know it. Nevertheless, as is the case with any other type of technology, the Ambient Mirror cannot be characterized, per se, as good or bad. Its use by our society will provide the final verdict.
In his afterword to Elhacedor ("The Maker"), Jorge Luis Borges wrote: "A man sets out to draw the world. As the years go by, he peoples a space with images of provinces, kingdoms, mountains, bays, ships, islands, fishes, rooms, instruments, stars, horses, and individuals. A short time before he dies, he discovers that the patient labyrinth of lines traces the lineaments of his own face ."
So, if in Borges's passage the man discovers in his drawing just "the lineaments of his own face," what would the same man discover by looking into the Ambient Mirror? The lineaments of his soul?
4. Zhai, S. and V. Bellotti. "Introduction to sensing-based interaction." ACM Transactions on Computer-Human Interaction (TOCHI) 12, no. 1 (March 2005):12. <http://doi.acm.org/10.1145/1057237.1057238>
Dimitris Grammenos is the lead interaction designer of the human computer interaction laboratory of ICS-FORTH, where he heads the lab's universally accessible games activity. He holds a B.Sc. in computer science and an M.Sc. and Ph.D. in electronic engineering. Grammenos has expertise on interaction design issues in traditional Windows-based environments, as well as in technologically advanced environments, such as virtual and augmented environments, wearable computers, and ambient intelligence environments. He has been involved in several European R&D projects related to design for all and universal access, and has given related lectures, seminars, and tutorials. For more information visit: http://www.ics.forth.gr/hci/people/dgrammenos.html.
©2009 ACM 1072-5220/09/0300 $5.00
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2009 ACM, Inc.