The European Commission made ‘ambient intelligence’ a focus of its research programmes for 2001 to 2005. In official documents, the commission sometimes replaces the words ambient intelligence with the acronym AmI – to which I, as a philosophical joke, started adding a question mark – as in,
Am I?
We shall soon find out: several hundred million euros have been earmarked for the design of systems composed of autonomous entities whose participation in computation is dynamic and where activity is not centrally coordinated.

Another action line is about integrated perception systems inspired by living systems – where ‘perception’ is meant to include sensorial, cognitive and control aspects. All of which reminds me of Bruce Sterling’s question at Doors 6: “What happens when you go to look at the flowers in the garden – and the flowers look at you?” Anyway, the following text is a first attempt to derive some design issues from this story. Comments welcome, as usual. (Note: this text draws heavily on the smart materials seminars we ran at the Netherlands Design Institute in 1998 and 1999: thanks to the colleagues concerned).
We tend to think of products as lumps of dead matter: inert; passive; dumb. But products are becoming lively, active, and intelligent. Objects that are sensitive to their environment, act with some intelligence, and talk to each other, are changing the basic phenomenology of products – the way they exist in the world. The result is to undermine long-standing design principles. “Form follows function” made sense when products were designed for a specific task – but not when responsive materials, that modify a product’s behaviour, are available. Another nostrum, ‘truth to materials’, was a moral imperative of the modern movement in design; it made sense when products were made of ‘found’ or natural materials whose properties were pre-determined. But ‘truth’ is less helpful as a design principle when the performance and behaviour of materials can be specified in advance.
‘Smart matter’ is a loosely defined category of physical materials which are combined with digital systems to create programmable matter that can change in shape, stiffness, colour, reflectivity and even sound. Also known as responsive technologies, smart matter comes in a variety of forms – ranging from wires and gels to inks and computers. Smartness is manifesting itself in a range of environmentally responsive technologies. According to scientists at Xerox’s Palo Alto Research Center (PARC),for example, such materials could be used in the future to build skyscrapers with ‘smart’ structural columns that can change their physical properties.These columns could stiffen the building to resist high wind loads, but could also soften it, to help it ride out shockwaves from an earthquake.
The science of smart matter is in its infancy, but early experiments at PARC have been encouraging. For example, a series of strain sensors and piezo-ceramic actuators have been incorporated into a simple column. When this column is loaded, the actuators apply tiny pressures that effectively squeeze the column back into shape in the exact position where the strain gauges are showing imminent buckling failures. PARC has shown that the buckling load of such a ‘smart’ column can be improved by a factor of two without adding any extra material. However such ‘active systems’ require very fast real-time computer power, and scaling them up to the deployment of millions of sensors poses great problems for any exisitng software control architecture.
Of penguins, polar bears and the feet of ducks
Nature, too, offers us countless examples of how to revolutionise the products, systems and structures that surround us.Biomimicry studies nature’s models, and then imitates or takes inspiration from these designs to solve human problems – for example, a solar cell inspired by a leaf. A pioneer in the field, the biologist Julien Vincent (Director of the Centre for Biomimetics at Reading University), looks at organisms as bags of answers and tries to figure out: what were the questions? What are the problems that nature was trying to solve? “That is really what biology is all about. Not just deciding what the problems were, but the way in which they have been solved and how these optimisations occur. Unless you understand the optimisations, you do not really understand the nature of the solutions.”
In Vincent’s view, the same goes for engineering. The similarities between between biology and engineering are immense. With any biological organism, you have a limited amount of energy. The energy comes from sunlight – or it is stored energy from previous aeons of sunlight – but all biological organisms are in competition with each other for the use of this energy. The most successful organisms take the minimum amount of energy and optimise the distribution of the energy between all its different functions. If it uses more energy than its neighbour, it finds it cannot reproduce so well, and does not leave so many offspring. It eventually dies out. “There are good reasons, based on evolutionary selection pressure, for reckoning that nearly all biological organisms represent minimum-energy solutions to particular problems,” says Vincent. That in itself is useful.
A novel form of technology and knowledge transfer, biomomicry is transforming how we invent, compute, heal ourselves, harness energy, do business, feed the world. But a conscious effort is needed – a design effort – to connect the properties of natural things to the needs we have as humans. When this happens, the results can be startling, profound and almost instant. But such connections do not make themselves.
It was in this context that the Netherlands Design Institute organised its first
workshop on smart materials. The idea was to bring together minds from very diverse areas, and participants were invited from aerospace; architecture; civil engineering; cognitive science; fashion; medicine; nanotechnology; product design and robotics. It was hoped that broadening the field of interaction of those involved might deliver the first steps towards a coherent agenda for designers working with these new technologies. The workshop was led by materials specialist Marie O’Mahony, a pioneer in the search for applications of ‘smart’ materials who was already actively looking at ways textiles would merge in clothing, human bodies, and buildings.
Current building technology is crude, and lacks functional integration. But new solutions may be found for insulation and heat storage by studying polar bears, penguins, cuttlefish, and even the feet of ducks.
Take the foundations that houses, factories and other structures rest on. The earth itself is easily strong enough to support houses, towers and offices. But builders have become accustomed to the tradition of hammering or pushing heavy concrete piles, weighing thousands of tons, deep in to the ground. A possible alternative is to build the structure on a light ‘floating’ raft made out of polystyrene foam. This can only be part of the solution since the density of the ground and the mud under a building tend to vary through time. Therefore the raft needs to be put on some kind of smart stabilising base that would be able to compensate for these changes. A structure that could do this could be composed according to the principle of a cuttlefish skeleton. This is extremely rigid although its volume is only seven per cent solid. Inside are channels that alternate with layers of plates. Inside these narrow channels is a gas which can make the fish go up anbd down by changing the pressure inside. If the pressure is low, the skeleton’s density will become relatively high, and vice versa. A cuttefish-style trim layer under a building could work in a similar way.
Or take insulation. Even though half the energy in the world is consumed in and by buildings, architects and building designers have hardly even begun to tackle this profligate and unsustainable performance; on the contrary, many architects in their search for transparency and the appearance of lightness in buildings have greatly increased the heaviness of their environmental footprint. Traditional buildings perform little better: in the classic sealed skin solution, a layer of bricks or cladding on the outside is backed by a layer of insulating foam: the system excludes the cold to a degree – but not nearly as well as the skin of a polar bear, which can maintain its body heat in the extreme cold.
The bear’s skin itself is black. It is covered in a thick layer of translucent white hairs; these combine with trapped air to form an insulating layer. It absorbs every little bit of heat brilliantly. The hairs themselves guide infrared light towards the skin. Each individual hair is able to convey any external heat back to the skin, which absorbs it. Transferring this kind of functionality to a building would mean a black façade: panels would consist of two transparent layers with transparent capillary-like tubes mimicking the hairs between them.
In any situation where temperatures change a lot, the example of a penguin might be more effective. In extreme cold, this amazing bird can stand on ice for weeks on end, and sustain a temperature differential between its own body and the outside environment of 80 degrees celsius. Even better: it can swim in icy water and then clamber onto a sunny beach without over-heating. One of the penguin’s secrets lies in its dense feathers. At one moment, these reduce airflow and behave like treacle, at another they open like valves, moving freely for ventilation purposes. Another secret is the way its feet are used to get rid of excess heat when it lands: blood vessels in its feet act like radiators, heating up the ground. A penguin-inspired building could do more.
I have a hammer, but I need a nail
Smartness, it is true, remains more of a promise than a reality in most industries. Many of the more exotic smart materials are still at experimental stage, and have yet to find widepsread applications. The tendency is for scientists and engineers to develop exotic performances in new materials – but to leave the development of useful applications to others. “We have the hammer, but we’re still looking for the nail,” say researchers. And thousands of products are still produced that contain no microchips and are dumb as dumb can be. According to Julien Vincent, there are three problems in translating ideas from biology into technology: “interpretation, implementation, and scale.”
But the connectivity part of the equation is growing exponentially. Microchips are ubiquitous: 200 billion are in use today, installed in everything from 747s to greetings cards. That’s 35 chips for every man, woman and child on the planet. Each of those chips has the potential, using existing technology, to be connected to the others. What happens then is that products become services. In these new, responsive, connected environments, the basic business logic – the source of added value – is the control of information. As more and more devices connect with each other in the factory and the home, and with telecommunication networks, value tends to be created by the supply and operation of networked devices – not in the hardware per se.
Everyone will be pushed remorcelessly into the information business. Builders, who once built ‘houses’, must now contend with issues of connectivity. Building suppliers, who once produced taps and doors and thermostats, must now turn them into smart devices. Consumer electronics companies, happy to sell you a dumb stereo in a box, must now teach it to talk to your television. Kitchen suppliers, accustomed to a world of chipbaord cabinets and counter tops, must contend with exotic materials and digital devices that change when heated or spoken to. Smartness is impinging on gas, electricity and water utilities, even health and welfare services.
Put another way: the pressure on businesses to consume less matter, and manage more information, is inexorable. These changes are already happening, but in a subtle way. Companies that already use smartness in commercially-available products – their products range from artificial legs, to fish-counting devices, and digital ticket systems – didn’t set out to make their products ‘smart’. But in solving real-world challenges, they realised that the availability of microprocessors, or new materials, made their task easier. As is often the case in the real world, innovation is a natural byproduct of running a business, not an end in itself.
The Blatchford company, for example, uses new materials and chips in its products – not for their own sake, but to help people walk. The British company, the first to develop an intelligent prosthetic limb, needed to make its products light, strong and easily formed into complex shapes – so it sought out and applied advanced carbon fibre composite materials to the task. It also needed to control, automatically, the flexion and extension of the limb to make walking easier, so it incorporated sensors and microchips.
The materials of invention
In exploiting the new opportunities presented by chips and new materials, companies are spoilt for choice. For the first million years or so after our appearance, we humans used five basic materials to make tools and objects: wood, rock, bone, horn and leather. For 9,000 years following the Neolithic revolution, there was a significant enrichment: clay, wool, plant fibres and, in relatively recent times, metals. But it still took at least a generation for information about smelting to cross a single continent in the Iron Age.
Today, in contrast, novel substances are being cooked up in academic and corporate laboratories faster than industry can find uses for them. An important book by Ezio Manzini, The Material of Invention, puts into perspective the revolutionary implications of new materials for product development and design. Designers and manufacturers are faced with an enormous and expanding field of possibilities – in the selection of materials, and of industrial processes to transform them. Known and trusted physical limits, that are deeply embedded in the skills and cultures of artisans and production engineers, are suddenly disappearing.
As well as choosing from numerous alternative ways to meet existing needs, designers also confront the problem: what should a product look like? What form should it take? New materials have no absolute form. Neither do they have ‘natural properties’, to determine a product’s shape. As Professor Manzini puts it, “a material of invention is no longer a found material; rather, one is calculated and engineered to achieve a specific, desired performance.”
The problem for companies and designers is that inventors categorise their materials according to what they are – rather than what they do, or what they are for. Grouping materials by type makes sense in an archive, but it does not help the product designer confronted by databases and directories bulging with thousands of plastics, ceramics, fibres, composites, rubber and foam, glass, wood, and metals. What’s needed are information systems that direct designers first to properties , and thence to the different materials that possess them.
Ironically, the most interesting property of new materials is their capacity to promote the dematerialisation of products – that is, the use of less matter to accomplish a given task. SRI, an American think-tank, evaluates new materials and processes against this benchmark. Researchers at Domus Academy, a design think-tank in Milan, agree: there is serious demand for smart materials from manufacturers striving to impart lightness, simplicity, and connectivity to their products. They give the following examples:
Lightness: a range of composites, laminates and so-called ‘structurally gradient’ materials are being used to replace heavier materials such as steel or concrete. Some new materials are so light and ethereal that we have not yet worked out how to exploit them: slicia aerogel, for example, is an ultralight material, comprising 99.9 per cent air, developed at the Livermore National Laboratories in California. One reviewer, Julie Wosk, said the substance ‘had the translucency of clouds and the eerie, phantasmagoric look of a hologram’. Aerogel is used as an insulating material, and as a filter, but has not yet been exploited by product design.
Simplicity: delivers quality – a consumer need which dominated the business agenda of the 1980s. There is a trend in all kinds of products to minimise complex and therefore failure-prone assemblies of sub-components. New materials that behave as if they were devices, but are not, enable the substitution of mechanical with membrane keyboards, or of lightbulbs with electro-luminscent surfaces.
Hardness: modern ceramics and advanced composites can be nearly as hard as diamonds and are equally resistant to heat and corrosion – but are cheaper. They are being made into turbines, dental braces, prosthetic body implants, even bullet-proof face masks.
Transparency: dematerialisation does not just mean less matter in a product – it also means less physical presence. Architects, who are leading the way in a search for ‘de-massified’ buildings, have stimulated the development of new stong but lightweight glazing systems, translucent panels, and light-reflective finishes that make objects disappear.
Environmental sensitivity: fibre optic sensors, which have been used for many years to deliver information, are now being combined with sensing technologies such as piezo-electric materials – shape memory alloys which change their shape according to temperature, so that distant environments may be monitored remotely. Applications range from security to medicine.
Nice to meet you
We are accustomed to conversing with a computer, interacting with an ATM. But if they have chips, and most of them do, our microwaves, fax machines, CD players, washing machines, light switches, thermostats, burglar alarms and door locks, also have the capacity to communicate with each other. The growing use of mechatronics – the intelligent electronic control and connection to each other of all manner of devices ranging from cars to pop-up toasters that now exist alone on our homes and workplaces – does not mean that dumb toasters will suddenly become philosophers. But it does mean that the borders between ‘products’ and ‘information’ are starting to dissolve. This is clear enough in the case of the Austrian company SkiData. Ostensibly in the business of dispensing millions of electronic tickets to people attending trade fairs, joining toll-roads, parking in garages, or heading for ski slopes, the Austrian company is in fact in the business of information network services; its hardware, and in particular its wrist-worn devices developed with Swatch – are what you notice, but in value terms they are simply interfaces to the system.
Most of us probably assume that microchips are small things that sit inside desktop computers – but 90 per cent of all microprocessors in use worldwide are in automobiles and ordinary domestic appliances. (Or in the case of an Icelandic company, Vaki, in boxes underwater counting fish). It famously takes more computing power to run a BMW today than it took to send early astronauts round the moon. The average value of electronic components in cars today is already $920; according to the Economist Intelligence Unit (EIU), that will soar to over $1,700 per car within a decade. “The world market for automotive electronics will be a staggering $83 billion by 2005,” says the EIU. “Electronics will become the driving force behind vehicle design.”
Where cars lead, other consumer products invariably follow. A computer company, Novell, estimates that there are about 145 microprocessor-controlled devices in every (presumably American) home. Novell is developing home-based systems, which it calls NEST (NetWare Embedded Systems), that permit any device to communicate with any other. When such systems are in place, a central computer can do everything from turning off lights to turning sprinklers on.
The potential for energy savings alone, when such networks combine smart sensors with an intelligent controller, are immense. Buildings consume some 50 per cent of the world’s energy – and by some estimates, at least half that could be saved if heating (and cooling) systems were more sensitive to the minute-by-minute needs of users. At present, even the most advanced buildings consume vast amounts of energy to sustain average temperatures, even when rooms or buildings are empty. Holmhed, a Swedish company that makes small demolition robots, may be part of the answer: its ‘jackhammers on tracks’ now have programmable remote controls – so presumably they could be instructed to go forth and tear down energy-inefficient buildings.
But microchips will not just enhance the practical aspects of running a house. Affective and aesthetical qualities can also be enriched. Artemide, the Italian lighting company, is legendary in design circles for totemic lighting fixtures such as the Tizio desktop lamp that graces designer desktops in countless films and advertisements and sells several hundred thousand units a year. But even Artemide, once a byword for object fetishism, is going digital with its remarkable Metamorfosi, which affords the householder up to twelve million different lighting combinations. So far, Metamorfosi is microprocessor controlled but stand-alone; but for how long? Potentially, it can now talk to the other lighting systems, so sooner or later it will.
A survey in the Economist estimated that 1,000 manufacturers are developing systems infastructures for the control of a houseful of gadgets – the software, transceivers, routers and other networking equipment that link the microprocessors in discrete products together. This technology takes the decentralisation of computing – which has its roots in the transition from mainframes to PCs – to an extreme, says The Economist. “Each chip has enough intelligence to carry out its basic task and to communicate its status to other chips on the network. To add intelligence to the system you (just) add a PC.” Once these smart sensors and microprocessors are joined together you get yet another network – and a context in which hardware and software become the infrastructure for a new kind of service.
Embedded, but not asleep
A world in which products and appliances talk to each other sounds fantastic, but it’s not so long since the advent of electricity was greeted with similar amazement: then, objects which once had to be worked by hand, began to power themelves. Soon, objects which we now have to control ourselves, will tell themselves and each other what to do. If the rapid electrification of everyday life just two or three generations ago life is any guide, embedded computing may not be so hard to get used to for consumers. But the task for product developers and designers is harder. How should a radically new technology (or rather, family of technologies) be introduced to a public that is either ignorant of the benefits, or downright hostile to the whole idea?
When electricity was first introduced into the home, there was a tendency in industry to portray its aims, technological prowess, and its dynamic power, in mythological terms. AEG, for example, used the goddess of light as its trademark. But once electricity’s magical novelty wore off, end everyday products began to be ‘electrified’, designers had to find new ways to make products such as electric irons, kettles, lightbulbs and cookers interesting to consumers. Reasoning that, “even an electric motor must look like a birthday present,” artists like AEG’s Peter Behrens turned themselves into industrial designers to accomplish just that. Today, electric motor technology has disappeared from view almost entirely. It’s there, humming away inside a swarm of everyday household products. Today’s network visionaries anticipate a future in which products are transformed from discrete lumps of matter into components of an integrated consumer appliance and office automation ‘medium.’
It’s a bewitching vision of the future – but one that may be further away than we think. James Woudhuysen, a former manager of long-term market research for Philips Sound and Vision, thinks these scenarios are suspect: “The consumer of digital media (and appliances) meets convergence only in the same way that, in a blizzard, the snow seems to ‘converge’ upon you. We all may hope that everything will one day be controlled by a single, submissive black box below the stairs – but in new technology, systems are more prone to being incompatible than to matching up with each other.” In contrast to the heady days when electricity arrived, the main task of design in connection with embedded intelligence is systems integration. The creation of new design languages to articulate the wonders of domestic smartness will be less of a priority for the time being.
Nonetheless, the nature of innovation and design will have to change. In 1903, AEG’s Paul Westheim observed of design at the dawn of the electrical age, “in order to make a lucid, logical and clearly articulated entity out of an arc lamp, a complete transformation of our aesthetic notions was necessary.” The same could be said of information technology today. A new presence has come into our lives, yet it lacks visible form or expression. As with electricity in 1900, so with embedded information systems now. Design is faced with an aesthetic, as well as practical, challenge: how to represent something essentially abstract in such a way that it can be used and appreciated by us all.
Ezio Manzini observes that we are living in the ‘twilight of mechanics’, an age of boundless possibility in which anything thinkable becomes possible – but when we do not know what ‘needs’ this explosion of material creativity is supposed to meet. As Manzini says, “The uncontrolled increase of performances and forms, made possible by technique, takes place beyond adequate cultural control, thus producing noise and trash.”