are we being experimented on? how would we find out?
Promises and Threats
Ambient intelligence is not science fiction, but actual reality that will, bit by bit, pervade all aspects of our lives. We are already used to shop doors that slide open automatically as we walk in, and detection systems that keep an eye out for fire. The step to more comprehensive intelligent environments is only a small one. Typical for such environments is their interactivity, as well as their total or partial invisibility. Intelligent environments consist of a continually communicating network of devices that are permanently in contact with the environment and respond to it actively and on their own initiative. This contact may be realized by microphones, cameras, infrared sensors or scanners that can read RFID chips2 without having to be connected to them by wires.
In their evaluation of ambient intelligence, the Information Society Technologies Advisory Group (ISTAG)—the most important advisory body of the European Commission for policy in the field of ICT—focused on the increased usability that often is the result of ambient intelligence. It gives users more influence, prevents complicated interaction with technology and enables more efficient services (, 1). The way Aarts and Marzano elaborate the concept of ambient intelligence, though—which broadly describes the Philips approach—does not understand ambient intelligence so much in technological terms as in social terms ([1, 2] cf. ). For them, it is important that a technological system has a form of social intelligence: it should be capable of intelligent interaction with users.
Aarts and Marzano distinguish five layers in this interaction, each of which builds on the previous layers. Firstly there is the layer of embedding: ambient intelligence is embedded in the environment, both in the physical sense, hidden in walls, clothing and packaging, and in the social sense: it is possible to communicate with it in a ‘natural’ way, for example by means of movement or speech. Secondly, this technology is aware of its environment: it responds to what happens around it, by detecting movement, reading RFID chips, recognizing speech and so on. Shops can, for instance, automate the payment and stock system extensively if all products are equipped with RFID chips that can be read automatically by the checkout. The third layer is that of personalization. Ambient intelligence can draw up or retrieve a person’s profile and set up interaction with technology that is tailor-made to suit the person in question. An example is a refrigerator that forms a picture of someone’s eating pattern and subsequently makes suggestions for the shopping list, possibly with dietary advice thrown in. Fourthly, this technology has the capacity to adjust; it not only detects its environment, but adjusts to it in a personalized manner. The intelligent refrigerator in the example might quite well be able to harmonize the menu suggestions to the time of year. The fifth and last layer in ambient intelligence is that of anticipation. This technology can not only respond to its environment, but can also think ahead, like a car that can anticipate the movements of other road users and adapt its speed automatically if other road users suddenly brake, accelerate or switch lanes.
Persuasive technology adds yet another step to these possibilities. Here intelligent systems and environments are explicitly deployed to influence human behaviour, to persuade people to act in a particular way. The art of persuasion has been recognized for millennia. From the rhetoricians and sophists in ancient Greece to the spin doctors and advertisers of today, people have developed technologies for persuading others of particular standpoints, to carry out certain actions or, on the contrary, not to do so. In the twentieth century, the art of persuasion became the object of behavioural scientific research. Not only is the form of the message now used to influence human behaviour, but also the characteristics of the receiver. By combining an understanding of how influencing behaviour works with the specific possibilities provided by information and communication technology, new leeway has arisen for the design and application of technologies that encroach greatly on our everyday activities and choice processes, and even on our ethical decision making (cf. ).
The FoodPhone, to give an example, is a specific application of mobile telephones with built-in cameras that are supposed to help obese people lose weight. If you take a photograph of everything you eat and send it to a central number, you receive detailed feedback on the number of calories you have eaten, so that you can relate that to your calorie consumption throughout the day. The Baby Think It Over is a doll that can be used in educational programmes to prevent teenage pregnancies. The doll gives a realistic picture of the amount of care and attention a new-born baby needs throughout the day and night, and tries in this way to motivate teenagers, from the inside, not to become pregnant at too young an age. The Persuasive Mirror mentioned earlier gives feedback on the health implications of recent behaviour by extrapolating someone’s mirror image to the future.
There is no doubt about the fact that the social impact of ambient intelligence and persuasive technology will be enormous . The possible futures relating to these technologies are both utopian and dystopian, as is the case with most technologies . On the one hand the promise is that ambient intelligence will remove itself to the background so that humans again become central. Because of its interactive character we should, moreover, finally have a technology at our disposal that adjusts to suit humans, instead of us having to adjust to the technology. In the meantime, we seem to be surrounded by beneficial applications: cameras that automatically monitor safety in public spaces, the automatic administration of medication in hospitals, technologies that help us find a healthy lifestyle, safety provisions in elderly people’s homes to enable them to live there longer. Provided well programmed, these technologies promise us a glorious new world.
On the other hand, there are dangers implicit in these technologies. Because of their interactive character, ambient intelligence technologies collect a lot of information about their users, and consequently form a new type of threat to our privacy. Furthermore, these technologies undertake responsibilities that have so far been the domain of people, and this does not necessarily take place safely and reliably. These are, however, aspects that apply to practically every new technology. More fundamentally, and more specific to ambient intelligence and persuasive technology, is the question of what happens to human freedom and responsibility here. If our environment starts to respond intelligently to us and to take our decisions for us, aren’t people going to gradually lose control over their own lives? Will we still take responsibility and be held responsible for our deeds? Are there still ways out of the controlling environments described? And what happens if persuasive technology consciously starts to influence our moral considerations? Do we want technology educating us? And who will then be responsible for the content of this education?