Philips and Palo Alto Ventures (1998) coin the term ‘ambient intelligence’ (AmI).This concept has been developed further and has contributed to products such as the Philips Healthcare Ambient Experience (de Ruyter & Aarts 2004). The concept of AmI grew bigger and received more attention as the European Commission’s Sixth Framework in Information, Society and Technology (IST) programme agreed to fund Ami related research projects (Aisola, 2005). It is possible that the funding was given provided they followed the advice given by the Information Society and Technology Advisory Group (ISTAG). ISAGT is a group that advises the European Commission on how to strategise and implement ICT research. ISTAG carried out an exercise and showed what it would be like to live with AmI in 2010 and have extended it to 2020. In addition to this the USA, Canada, Spain, France and the Netherlands have also carried out research on AmI. The first conference of AmI was held in 2004 and many other symposiums have set up regarding AmI since then.
There are other technological developments that are similar to AmI for example Artificial Intelligence (Aarts, 2003). AmI is also closely related to ubiquitous computing and aspects of AmI were already covered in the early 1980’s/early 1990’s and this was the work of Marc Weiser. Marc Weiser vision has almost been achieved especially when we look at the progress of AmI to date. There is a big difference between the early concept of AmI and the current concept. It is believed that there has been a shift in AmI in the sense that earlier visions were more concerned with productivity in business environments where as the current vision is concerned with consumer and user centred design approach (Aarts ,2003).
A common example of AmI is a fridge that orders products by itself without the help of humans (Ducatel, 2010).The fridge has the ability to analyse consuming behavior patterns for example, there is always milk in the fridge and it is programmed to order milk when there is only one bottle left. A Miniaturized biosensor systems is another example of AmI. A miniaturized biosensor system monitors various bodily activities such as blood pressure, glucose levels, body temperature, heart rate etc. The biosensor systems are connected to an emergency unit that automatically sends an ambulance when there is an accident. In this case it automatically ‘assumes’ when certain bodily variable values have exceed the average (Schuurman et al., 2009; Federal Ministry of Education and Research, 2007).
Technology that uses AmI are often connected wirelessly to create intelligent networks that allows society to interact intelligent and intuitive interfaces. Holmlid & Björklind (2003) believed that the ultimate goal is to design an environment that is capable of recognizing and responding to the presence and actions of different individuals in a seamless, unobtrusive and often, invisible way using several sensors.
Definition and Defining Features
AmI in electronic devices is usually embedded, interconnected, adaptive, personalized, anticipatory and context-aware. These are the six defining features of AmI and will be discussed bellow.
– Embedded: AmI is not the result of technology or an application but more of a property of the electronic device. Such Ami devices are embedded in the surroundings without people consciously noticing.
-Interconnected: As well as being embedded in the environmentdevices, sensors and ICT systems are also wirelessly interconnected and this creates an ubiquitous system. An example can be drawn from the miniaturized biosensor systems. The miniaturized biosensor systems monitors variables in our body and this is connected to an emergency unit which is linked to the ambulance. All the connections made between the devices, sensors and ICT systems are wireless connections and make an ambient intelligence system.
-Adaptive: AmI systems are very adaptive in the sense that they do not rely on up to date and complete information to function. AmI has no stable connectivity or source to information. For these reasons the service of the AmI system may vary depending on the accessible information and the reach-ability of external services.
– Personalized: AmI can be personalized to fit the users preferences. An example of this is the interactive interface in a mirror which informs the user of the weather, traffic updates and appointment reminders. One could say that AmI is user-centered.
-Anticipatory: AmI is also anticipatory as it can predict its user’s desires. For instance say there is an AmI system at a person’s home that monitors their behavioral pattern, it then comes to a conclusion about their current mood and adjusts the light and music accordingly. This is in attempt to anticipate the user’s needs so again one could say it is user centred.
– Context-aware: AmI technology has the ability to distinguish users and their situational- context and can therefore change the user and context. Such systems may know the type of music their user prefers or their preferred house temperature. A common example of context-aware AmI device is car navigation. The system automatically adjusts the screen lighting level when it gets darker and in this context the user would benefit from this action.
– Novel human-technology interaction paradigms:Advancement of AmI includes new types of interfaces that use interaction paradigms like speech or haptics.
All defining features of AmI differ in presence in all AmI systems. There are specific AmI systems that are very personalized, anticipatory and adaptive and there are other AmI systems that are less concerned with these features. In addition there are some overlap between a few of the defining features including adaptive, personalized, anticipatory and context-aware. The fact that AmI systems are adaptive and context-aware adds to the personalization aspect for the user.
– AmI in the home environment: ‘Ellen returns home after a long day’s work. At the front door she is recognized by an intelligent surveillance camera, the door alarm is switched off, and the door unlocks and opens. When she enters the hall the house map indicates that her husband Peter is at an art fair in Paris, and that her daughter Charlotte is in the children’s playroom, where she is playing with an interactive screen. The remote children surveillance service is notified that she is at home, and subsequently the on-line connection is switched off. When she enters the kitchen the family memo frame lights up to indicate that there are new messages. The shopping list that has been composed needs confirmation before it is sent to the supermarket for delivery. There is also a message notifying that the home information system has found new information on the semantic Web about economic holiday cottages with sea sight in Spain. She briefly connects to the playroom to say hello to Charlotte, and her video picture automatically appears on the flat screen that is currently used by Charlotte. Next, she connects to Peter at the art fair in Paris. He shows her through his contact lens camera some of the sculptures he intends to buy, and she confirms his choice. In the mean time she selects one of the displayed menus that indicate what can be prepared with the food that is currently available from the pantry and the refrigerator. Next, she switches to the video on demand channel to watch the latest news program. Through the follow me, she switches over to the flat screen in the bedroom where she is going to have her personalized workout session. Later that evening, after Peter has returned home, they are chatting with a friend in the living room with their personalized ambient lighting switched on. They watch the virtual presenter that informs them about the programs and the information that have been recorded by the home storage server earlier that day’ (Wikipedia, 2010).
AmI in healthcare: ‘Ambient Intelligence could provide a basis for integrating intelligent health care technology into an individual’s personal surroundings. Computers around you, on your body and even in your body could monitor your health status at all times and, when the need arose, alert your carer or intervene directly’ (Schuurman, El-Hadidy, Krom & Walhout, 2009, p. 15).
– AmI for disabled persons: ‘A personal communication device can be worn or fitted to a wheelchair or a blind person’s cane. These can be programmed to communicate with barriers, ticket machines and gates to allow access or more time. Smart tags, embedded in a floor, can receive and send information that will guide a person to a destination. A person with low-vision could hear guidance signals’ (Gill, 2008, p 6).
– AmI for industry: ‘Intelligent and autonomous networked sensors systems offer possible uses among others in precise, low-cost control of chemical processes, in monitoring and linking machines, in the tracking and management of security-relevant objects, in registering ambient conditions and in testing product quality of the condition of building fabric’ (Federal Ministry of Education and Research, 2007, p 39).
– AmI for business: ‘Hélène calls Ralph in New York from their company’s home office in Paris. Ralph’s E21, connected to his phone, recognizes Hélène’s telephone number; it answers in her native French, reports that Ralph is away on vacation, and asks if her call is urgent. The E21’s multilingual speech and automation systems, which Ralph has scripted to handle urgent calls from people such as Hélène, recognize the word “décisif” in Hélène’s reply and transfer the call to Ralph’s H21 in his hotel. When Ralph speaks with Hélène, he decides to bring George, now at home in London, into the conversation. All three decide to meet next week in Paris. Conversing with their E21s, they ask their automated calendars to compare their schedules and check the availability of flights from New York and London to Paris. Next Tuesday at 11am looks good. All three say “OK” and their automation systems make the necessary reservations. Ralph and George arrive at Paris headquarters. At the front desk, they pick up H21s, which recognize their faces and connect to their E21s in New York and London. Ralph asks his H21 where they can find Hélène. It tells them she’s across the street, and it provides an indoor/outdoor navigation system to guide them to her. George asks his H21 for “last week’s technical drawings,” which he forgot to bring. The H21 finds and fetches the drawings just as they meet Hélène’ (MIT Project Oxygen, 2010).
The idea of AmI is a vision of the future advancements in ICT and is actually not available as of yet, especially not the complex definitions and examples that have been mentioned in this case study. Common devices are to some extent context-aware, personalized and adaptive but nowhere near the type of level of the AmI vision. 2020 is the year predicted by Philips Research (2010), Gasson & Warwick (2007) and Federal Ministry of Education and Research (2007)that AmI will be readily available to users.
Relation to other Technologies
As mentioned before AmI is linked to other ICT concepts for instance ubiquitous computing and the internet of things. These concepts are very similar to AmI and often used as alternatives. However it could be argued that there is a difference between ubiquitous computing and the internet of things and AmI in the sense that ubiquitous computing and the internet of things enable technologies for AmI. AmI would not be able to work without ubiquitous devices. Other technologies that are related to AmI include sensor systems, Radio Frequency Identification (RFID), interactive screens, televisions, radios, fridges, lighting systems and the Internet.
AmI is still quite away from coming in to existence and even when it finally does there is worry that it will be very difficult to maintain the infrastructure for AmI. The functional and non-functional requirements of AmI are extremely demanding on the computational and communication resources and on the underlying infrastructure (Gasson & Warwick, 2007). Gill (2008) emphasized that AmI systems must be able to ‘infer the goals of the users without giving them the impression that they are under control (big brother), and must be able to support the users without giving them the impression that they are forcing them. The system must offer possible solutions, not impose them. This requires a lot of ingenuity also on the part of human beings, and appears particularly difficult for a machine’ (p 5). In addition to this AmI is a very complex concept that incorporates cognitive and sensorial aspects and people’s ability to cope with the hyper stimulation and the corresponding cognitive load is unknown. Especially when considering individuals with reduced abilities and cognitive limitations (Gill, 2008). For these reasons there are technical issues and human-centered design issues that need to be further looked in to.
Below are a few critical issues that have been considered.
Data Protection, Surveillance, Privacy
Privacy is the most obvious ethical issue linked to AmI. Sensors are heavily involved in profiling and are needed to adapt to user preferences. AmI Systems monitor our behavior and this information is digitized, stored, retrieved and processed therefore invades an individual’s privacy. However questions have arisen in relation to the use of sensors for example, Who is in control of the personal data collected by the sensors and who has access to it? An operator? A hacker? The fact that AmI uses surveillance, in some cases, causes privacy issues. As such the user profiles need to be protected (Gasson & Warwick, 2007; Gill, 2008; Schuurman et al, 2009). Brey (2007) strongly believes that privacy is a right and that access to personal information should be controlled.
Autonomy, Trust, freedom
The concept of AmI is very smart as is adaptive, context-aware and anticipatory and can therefore make decisions for people and can influence a person’s autonomy (Brey, 2005). Autonomy, ‘is the ability to construct one’s own goals and values, and to have the freedom to make one’s own decisions and perform actions based on these decisions’ (Brey, 2005, p 94). Some believe that AmI allows humans to have more control however others argue otherwise. For example an AmI system can predict what their users like and order from an online store. In doing this users are free from carrying out minor routine tasks. Another example, to show that Ami systems can give users a sense of control, can be drawn from using interactive mirrors. Interactive mirrors can provide personalised information about the user’s environment and can inform them of the weather. If it is predicted to rain then the user knows to take an umbrella out.
However, these features may not always be ideal and can take away human control. Recall the fridge that automatically orders milk when there is one bottle left. The system has been programmed to do specific tasks and people rely on the system and trust that it does the tasks well. However what if on another occasion we do not want milk but orange juice? Who is to blame? The system or its user? Or what if the fridge orders water, but it should have ordered orange juice? This example illustrates how AmI can reduce one’s autonomy and perform tasks that do not coincide with the needs or intentions of its user (Schuurman et al, 2009; Gill, 2007). Another example is the use of AmI systems to adjust the light and music according to observations and inferences made about the users’ behavioral patterns. The system could easily misinterpret it’s observations because it is very difficult for a machine to predict human needs when humans themselves do not always know what they want. Søraker and Brey (2007) identified the drawback of AmI to be similar to that of behaviourism because AmI systems are often adaptive and anticipatory. Behaviourism, in a nut shell, claims that mental states can be inferred from behaviour. AmI has specific features that could make an individual feel that they are not in control of their environment.
‘One central challenge consists of achieving reliability and fail-safe stability in autonomous systems with ambient intelligence’ (Federal Ministry of Education and Research, 2007, p 39). A user relies on an AmI system to open the door when they stand in front of it stands in front of it. But if the door is closed and refuses to open the person cannot get through. We depend on the system and the system could experience some problems if the system fails or malfunction,
How much control do humans have over AmI systems? These systems could have harmful consequences and if they do then who is responsible for this? (Duffy, 2008 and Marino, 2006 ) These are a few of the questions that arise in relation to AmI and Responsibility. Humans do not have control of AmI systems due to specific defining features hence the reason why they cannot be held responsible for the use of such technology. This means that when AmI systems malfunctions users, developers and designers of the system cannot be fully accountable for the problem (Gill, 2008). This is known as the responsibility gap. According to Marino (2006) a person can be rightly responsible, “he knows the particular facts surrounding his action, and if he is able to freely form a decision to act, and to select one of a suitable set of available alternative actions based on these facts.”The responsibility gap can be bridged. Rather than considering the awareness and the ability to control consequences of AmI the identification of possible damages, social sustainability and how compensation for these damages is to distributed should take prominance (Marino, 2006). This in turn will address responsibility and liability problems.
There is a health gap between economically developed countries and the less economically developed countries as well as health gaps between the rich and poor in economically developed countries. AmI would be extremely useful when/ if applied to healthcare as such systems can monitor and predict patient’s needs more effectively than humans. However, this will encourage the gap between the economically developed countries and the less economically developed countries.
Attribution- This document was created by Jade Ijeh on the basis of research undertaken in the Ethical Project
Aisola, K. (2005) The European Union’s Sixth Framework Programme: a Laypersons Guide to Funding. Tactical Technology Collective, Amsterd
Brey, P. (2005). Freedom and privacy in ambient intelligence. In Proceedings of the sixth
International Conference of computer Ethics: Philosophical enquiry, Philip Brey, Frances Grodzinsky and Lucas Introna (Eds.). P. 91-100.
de Ruyter, B., Aarts, E. (2004): Ambient intelligence: visualizing the future, Proceedings of the Working Conference on Advanced Visual Interfaces, AVI 2004, pp 203-208
Duffy, B.R. (2008). Fundamental Issues in Affective Intelligent Social Machines, The Open ArtificiaIntelligence Journal, pp.21-34 (14), ISSN: 1874-0618 Volume 2, 2008
Federal Ministry of Education and Research (BMBF). (2007). ICT 2020 – Research for Innovations.
Gasson, M. & Warwick, K. (2007). D12.1 Study On Emerging AmI Technologies. FIDIS – Future of Identity in the Information Society
Gill, J. (2008). Ambient Intelligence – Paving the way. Cost.
Gill, S. P. (2008). Socio-ethics of interaction with intelligent interactive technologies. AI Soc. 22,(Jan. 2008), 283-300. DOI= http://dx.doi.org/10.1007/s00146-007-0145-y
Holmlid, S. & Björklind, A. (2003). Ambient Intelligence to go, AmiGo white paper on mobile intelligent ambience. Research Report SAR-03-03.
Marino D., Tamburrini G. (2006). Learning Robots and Human Responsibility, IRIE, InternationalReview of Information Ethics 6, pp. 46-51.
MIT Project Oxygen. (2010). Retrieved January 29 2014, from http://oxygen.lcs.mit.edu/
Philips Research. Retrieved January 30, 2014, from http://www.research.philips.com/technologies/projects/ambintel.html
Schuurman, J.G., El-Hadidy, F.M., Krom, A. & Walhout, B. (2009). Ambient Intelligence – Viable Future or Dangerous Illusion? Rathenau Institute, The Hague
Wikipedia entry on Ambient Intelligence. Retrieved February 2, 2014, from http://en.wikipedia.org/wiki/Ambient_intelligence#CITEREFAartsHarwigSchuurmans2001