Isabel Alvarez and Neil McBride
Introduction to Speckled Computing
Specks are minute semiconductor grains of around 1mm3 across which can carry out autonomous computing functions. They have their own power sources and will draw power from the environment. They also benefit from nano-scale sensor technology, so can measure selected effects in their environment, including temperature, movement or location. There is the promise of wide sensing ability, for example, of chemicals in the environment. Besides nano-sensors, specks have wireless communication and can receive and process data. A speck is hence autonomous, possessing its own battery, sensors, processing power and communication abilities.
Specks may then collaborate and communicate with other specks to form networks of specks – specknets – which can then be programmed to carry out large computational tasks. It should also be noted that besides the external programming of specknets, such networks will have the capacity to self-organise. These wireless sensor networks (WSNs) will have a wide range of applications, for example in environmental science, health, military and education.
Specks and specknets promise the ultimate ubiquitous computing. They can be sprayed into the atmosphere, onto buildings and onto people. A room with specks painted onto walls could provide a powerful computing environment where users could access the computing power of specknets wirelessly while present in the room.
Since each speck contains its own processes, communication and sensors, the specknets that form do not need base stations or any central control. They are truly autonomous and, depending on how they are programmed can self-organise to deal with complex problems and adapt to environmental needs. The Centre for Specked Computing promises that “Surfaces, walls, floors, ceilings, articles, and clothes, when sprayed with specks (or “speckled”), will be invested with a “computational aura” and sensitised post hoc as props for rich interactions with the computational resources.”
While the hardware technology for specks is progressing rapidly, software approaches to how to manage and program self-organising networks lags behind. Methods are being developed which apply the biological concepts of the immune system in order to enable self-organisation in specknets (Davoudani et al, 2007).
Applications are being developed which use small numbers of precursor specks, in larger formats. These specks may be attached to specific places on the person. For example, a distribution of specks supports gesture analysis such that, for example, the rotation of wrists can be monitored and the social acceptability of the resulting gestures considered.(Brewster, 2009) Specks can aid the monitoring of diseases such as chronic obstructive pulmonary disease. Attaching specks to the chest enables a more holistic monitoring of breathing patterns covering the whole chest, rather than attaching one or two sensors.(Rablinovich, . 2010). Specks may be used in the monitoring of the natural environment. For example, Specknets may allow the tracking of fundamental processes occurring in previously inaccessible aquatic environments, including understanding how oxidation-reduction occurs in lakes using a specknet of REDOX sensors (Spears et al, 2010). Specknets can also be used to evaluate the performance of occupied buildings Hence, most current applications of specknets, still in their infancy, concern the development of networked approaches to gathering sensor information which in then centrally processed. The role of the speck is then no more than an intelligent sensor, communicating its findings using mobile communications.
The Ethical Environment of Specknets
Specknets can be both pervasive and invisible. We could enter a room without being aware that a specknet was monitoring our presence, perhaps by each speck having a heat monitor and a detailed, holistic picture being built up of the environment. We may not be aware that specknets have been sprayed on our clothing such that social behaviour, breathing, or chemical changes in our bodies could be monitored, or that they might have been breathed in and monitoring us internally.
Specks’ ability as sensors might result in information about the human subject being collected without consent. While sensors are nothing new, specks provide added power and hence added possibilities in misuse in that whole populations can be used to gain much more detailed, holistic pictures of human behaviour in an environment than a few fixed sensors in a room or on a road might provide.
Issues concerning the participant’s awareness of the presence of a specknet, and consent to the recovery, analysis and interpretation of sensor data from specknets are clearly of importance. The participants’ autonomy and hence ability to make decisions about what data is release is compromised by the presence of a specknet. There is a requirement for transparency such that the presence of the specknet is clearly indicated and all specks can be accounted for. In such an environment this raises a need for specknet detectors which can reveal the presence of the specknet to participants. Like any ubiquitous computing the embedding of the computing power within the environment requires the provision of knowledge and access methods. This responsibility is clearly placed on the stakeholders who construct and deploy the specknets. The invisible nature of specknets raises issues about their usability which demand the development of interfaces which map the specknet, connect with the specknets and enable it to be interrogated as to both the processes it is performing and the data it is collecting and using. Also an ability to disable all specks in a specknet should be part of any specknet application.
However, unlike fixed ubiquitous computing system where the computing power may be embedded in the environment, perhaps hidden in the wall or as a laminate in a table, Specks have the potential for mobility and to be spread over a wide area. The problem here is that once specks are released into the environment, say through an aerosol spray, their locations cannot be fully controlled. The movement of participants, winds, drift to other garments or spread by animals means that control over the computing power and its location is reduced. An element of uncertainty is introduced into the ethical environment. The distribution of the specknet becomes uncertain and IT asset management of specks virtually impossible. Specks then have the potential to become environmental contaminants, possible interfering with other computing processes.
Does this mean that methods of containment are required? If we restrict specks to closed environments then a significant aspect of the specknet – its ability to spread in environments an monitor whole bodies in health applications or whole ecosystems in environmental applications is lost. Part of the value of specks lies in their mobility and penetration of the environment. Any ethics of specknets must balance the social value of the technology with the risks. A compassionate consideration of an environment or the health of an individual might suggest that the risks of distributing free agents such as specks is outweighed by the value in saving an environment or a life.
But what is also suggested that specks, once distributed remain the responsibility of the distributor. In extracting oil from a forest, the value to helping society and people thrive does not remove the empathic consideration of the environment. Reparation is required.
So ethical reflection suggests that the use of specks in an environment must be accompanied by the development of tools and technology that support their use. Computer simulations should be employed to predict their spread and locations. While there is uncertainty in distributing specks, say through an aerosol. Simulations of the environment should help predict patterns of dispersal. Such information would be essential anyway for the analysis of the environmental data provided by the speck’s nano-sensors. Additionally, technology needs to be developed which can be used in the field to monitor their presence and track their distribution. Also technology should be developed that can extract specks from environmental substrates and return them to the lab. Hence in ethical terms, while there is value in distributing specks and problems with negative ethics which constrains speck technology to fixed, contained environments, the development of supporting technology and the addressing of the communities within specks may be distributed demands the development of simulation, detection and recovery technology. The focus of the research cannot remain just on the specks themselves, but research must be cast in a context. Speck technology involves the community or environment within which they practice /their processing.
Specknets and Systems Ethics
While individual specks may be simple and controllable, their recruitment into networks will result in complex systems. A specknet is effectively a complex adaptive system. Simple rules and interactions at an individual level give rise to emergent behaviour of a higher order of complexity. In simple terms we can imagine flocking, in which simple behaviour by individuals gives rise to more complex group behaviour. A complex adaptive system not only displays emergent behaviour which is difficult to predict from the properties of individual specks, but may be dynamic and unstable.
Self-organisation, which could be envisaged in a specknet, may result in complex coordinated behaviour which includes the differentiation and specialisation of roles by groups of specks. Such complex adaptive behaviour is well documented in biology. For example, slime moulds are aggregations of single celled amoeba. These aggregations arise from simple behaviour in individual amoeba in response to chemical changes in the environment. The cells aggregate to form moving bodies which can differentiate to form specialised fruiting bodies. The slime mould can also exhibit other behaviours such as appearing to farm bacteria as a food source, carrying bacteria around and depositing them in more nutritious environments. Could specknets develop behaviours equivalent to slime moulds?
One problem is that such complex, dynamic , non-linear behaviour is difficult to predict from studies of individuals. And even if we adopt the approach of scientific determination, the complexity of the behaviour, responding to environments and involving a large number of variables will be difficult to model using a computer simulation. In many areas the complexity of dynamic changes in the environment adds to the potential unpredictability of the specknet. Furthermore, changes to specknets resulting from specialisation in purpose by groups may be irreversible.
Hence when specknets behave as complex adaptive system (which they have the potential to do so) in complex changing environments, there will be a substantial level of uncertainty as to what behaviour will occur and what the outcome will be. Preparedness through simulations and consideration of the options for behaviour will help, but at the end of the day, it is in the nature of complexity that unexpected behaviour will occur.
In ethical terms, the resulting ethical issues will be equally unexpected and sets of rules may be either too general or not adaptive enough. It is in the nature of the complex adaptive systems that ethical thinking will be done on the hoof and cannot be adequately served by codes of practice or rule-based systems. Virtuous approaches by the developers are required in which virtues such as empathy and patience are developed through learning and reflection. The role of the ethicist is, then, to encourage the developers and distributors of specks to maintain an ethical awareness, to reflect on the possibilities and to learn moral wisdom which will be adaptable to the outcomes of specknets.
The development of an ethical framework for specknets should be neither libertarian or luddite. The development of a system ethics is required.
Firstly a system ethics will involve the consideration of boundaries. There should be approaches to drawing specks in, to identifying the physical and logical boundaries which will define their sphere of practice. How can the specks be allocated to a physical boundary? How can the logic be defined which constrains them to the problem-in-hand and the practices needed to solve that problem?
Secondly, hierarchies should be considered. Emergent behaviour occurs at a higher level than that of individual specks. At these higher levels, new meanings emerge and detailed information is lost. Hence hierarchies of ethical behaviour may be developed. Understanding the level at which behaviour emerges will enable us to understanding of the ethical issues at that level to be developed.
Thirdly, an ethics of networks should be developed. Ethical issues emerge from interactions between specks within the networks. An understanding of the messages and relationships between specks within the specknet will lead to understanding of possible ethical interactions, since ethics is grounded in relationships and complex adaptive behaviour emerges from relationships.
Speckled computing offers a radically different approach to computing which may result in new emerging ethical problems. These problems go beyond the generally recognised set of problems.
New ethical problems arise from the potential for complex adaptive behaviour which may emerge in specknets. These ethical problems require a systems ethics which recognises the issue of uncertainty and unpredictability which may arise in specknets.
Ethical considerations can help identify technical needs associated with specknets including technology for monitoring and recovery. Additionally wise setting of boundaries can provide environments for creative and valuable use of specknets for human flourishing.
Briewster, S. (2010) Body-based gestures and social acceptability 9th Workshop of Specked Computing, Edinburgh, 24-25th November 2010
Centre for Speckled Computing (2009) http://www.specknet.org Accessed 17/03/11
Davoudani, D, Hart, E and Paechter,B. (2007) An immune-inspired approach to speckled computing. In DeCastro, L, Von Zuben, H.D. and Knidel, H. ICARIS 2007 Lecture Notes in Computer Science 4628 288-299
Gill, Z and Happlod, B. (2010) Evaluiting the performance of occupied buildings – an opportunity for specks? 9th Workshop of Specked Computing, Edinburgh, 24-25th November 2010
Rabonovicth,ER.A. (2010) Speckled Healthcare. 9th Workshop of Specked Computing, Edinburgh, 24-25th November 2010
Spears, B, Dudley, D and Harley J. (2010) The need for REDOX measurements in freshwater ecology. 9th Workshop of Specked Computing, Edinburgh, 24-25th November 2010
Sayed Chhattan Shah, Fida Hussain Chandio, Myong-Soon Park, “Speckled Computing: Evolution and Challenges,” icfn, pp.181-185, 2009 International Conference on Future Networks, 2009