The idea of “ubiquitous computing” emerged about 20 years ago. It promised great progress, but also stoked fears that we might fall prey to complete tutelage. Where are we at today?
Friedemann Mattern: From the very start, the notion of “ubiquitous computing” relied heavily on images that seemed very futuristic at the time. Certainly not all of these visions have become a reality. Several things have taken shape, however, including intelligent assistance systems such as the ones available in today’s smartphone apps. We could only dream of these kinds of assistance systems 15 or 20 years ago. At the time, it was also hard to imagine that these systems would be available on mobile phones equipped with internet access.

The buzzword “ubiquitous computing” sounds alarm bells with respect to privacy and the emergence of “the glass man” phenomenon. Have attitudes changed in this regard?
Even 30 years ago, people had great concerns about computerisation and privacy. Back then, the main concern was the introduction of PCs in the workplace and the concomitant transformation of work processes – something we have nearly forgotten today. What is happening today, of course, has taken on a new dimension: information technology is constantly becoming faster, smaller and cheaper. Sensor systems automatically generate data, and cameras can detect specific objects or faces with ever-greater precision. This has a much broader impact on our daily lives.

How?
On the one hand, privacy is now directly affected. This brings with it potential threats to the private sphere. On the other hand, we have changed our habits and behaviours, even simply through our use of smartphones, for example. We do things much more spontaneously, we don't have to make specific plans and we can react dynamically to situations. This is considered normal and a good thing. However, it also changes what systems know about us. An initial indication that pervasive computing has the potential to stimulate social upheavals was perhaps Google Street View. Now the next wave is approaching.

What shape will this take, and what will it entail?
Many of the assistance systems in question are supposed to improve our lives or make us safer. In order for these systems to provide users with information tailored specifically for them, however, the systems must have a certain amount of information about a person or must automatically obtain this information. One example that will be discussed at the conference is how we can manage energy efficiently. If we have detailed data about a household's energy consumption, it is possible to analyse and comparethis data automatically with similar households and determine why one household is consuming so much more energy than the other. We can then automatically generate specific tips on how to save energy effectively. Even something that sounds very positive, like this, does of course encroach on our privacy – often in an imperceptible way.

What happens if this information falls into the wrong hands?
The question as to whether this information can be abused is a serious one. Ubiquitous computing often generates large quantities of data that are processed for our benefit. When a sensor-based system observes me and learns when I typically arrive home in order to save heat energy, this is a good thing. But what if a burglar obtains this knowledge? It is clear that in the future, “smart” systems will have to be absolutely secure and trustworthy. However, I do think that some of the widely discussed threat scenarios are somewhat exaggerated. Does it make a significant difference whether my usage data is sent hourly to my power supplier via a secure communications channel or whether the meter is read manually once each month? Or that my neighbour might theoretically be able to hack into my “smart meter” via WLAN and obtain information about my power consumption – instead of simply going down to the basement and reading the meter?

Google Glass will also be discussed at the conference. If we always have to assume that we are being “monitored” and filmed by someone, how will this shape our lives in the future?
Interestingly enough, the technology used in these smart glasses is not significantly different from that of smartphones. But because you can always wear the glasses, many might feel uneasy or have legitimate concerns that their right to privacy is being violated. If this technology catches on after a few years, people probably won't want to do without these kinds of glasses because they offer a sense of security. In an emergency, for example, an emergency doctor or the police will be able to access footage from the glasses remotely and see for themselves what is happening in real time – practically through the eyes of the wearer. I can also record anything. In an accident, I will be able to prove that I was not at fault because I have recorded the entire incident in a live stream. If people see such a tool – which is practically an extension to one's own senses – as providing security and as something valuable, it will be impossible to ban it completely.

What does that mean for society and our legal system? Under some circumstances, there will need to be accountability for things that went unpunished or could not be proven in the past.
The downside is that a person might be involuntarily recorded on a video stream of an unknown third party. However, this is a bit different from states using cameras for the surveillance of public spaces. If the glasses are permanently connected to the internet in the future, then theoretically everything I see can be stored. With or without data glasses, what is generally looming is that more and more data about us and the situations we experience can be stored, processed and used. This certainly poses a real challenge for our society!

If the smart glasses record how I am distracted while driving and crash into something, the evidence is against me. I no longer have the option of telling a white lie as an alibi.
You raise an important point. Up until now, there has at least implicitly been a right to forget. Now if everything is documented automatically and can be retrieved exactly, things can no longer be forgotten. This can sometimes be cruel. I hope our society and legal system will do a good job of dealing with this and give precedence to what is humane over what is technocratic. Generally, however, we must accept the fact that the internet is increasingly pushing its way into the physical world. This has enormous consequences – both positive and potentially negative ones.

Pervasive computing has already penetrated many spheres of life. Have you run out of things to work on?
Research in this area is sure to continue! On the one hand, this is because realising classic visions requires a lot of fine-tuning. On the other hand, research itself brings about new visions. Take smart glasses, for example. If you connect them with the appropriate means of communication, a bush doctor in an African village can link up with far-away specialists, allowing them to have a look at a particular case. This way, the doctor can examine the patient through the eyes of the bush doctor. There is enormous potential in the research of these kinds of assistance systems. “Tele-cars” are yet another vision waiting to be explored. They would make the driver redundant, as the car would be driven by someone in a central location using cameras. A console provides “force feedback”, which feels like real driving. From a technical standpoint, this is no longer science fiction and might appeal to taxi or car sharing companies.

What will we see at Ubicomp 2013?
One major topic is the automatic identification of situations, without people feeling burdened by the technology. New methods for pattern recognition and automatic learning are important for assistance systems. A “smart” system must be able to recognise, learn and abstract the typical patterns a person exhibits. These are difficult tasks.

What visions are practically ready for implementation?
There is a desire to offer to-the-meter location accuracy of people using smartphones in buildings such as airports or train stations. This location information would then be used to provide contextual information, which requires completely new technologies. This topic will also be covered at the Ubicomp conference. A long-cherished dream has also been the possibility of quickly locating lost objects. Of course, locating objects is nothing new; this technology has long been used in freight logistics to determine the location of valuable goods. But in the past, these systems were expensive and large. The driving force behind pervasive computing, however, is that everything is becoming cheaper, smaller and faster. If, in a few years, locator devices are only the size of a sugar cube and available for ten francs at the supermarket and allow the user to log into a website showing its location on a map, people will probably use it for all kinds of different things. You can just imagine the challenges these sorts of developments will pose for society!

About Friedemann Mattern

Friedemann Mattern is 58 years old and has been a professor of computer science at ETH Zurich since 1999. In autumn 2002, he founded the Institute for Pervasive Computing together with his colleagues. His research interests include concepts and models for distributed systems, ubiquitous computing, sensor networks and infrastructures for the Internet of Things.

Ubicomp 2013 at ETH Zurich

Ubicomp 2013 will be held at ETH Zurich from 8 – 12 September. The conference combines the major events in this field, including the Pervasive conference (which was launched in Zurich in 2002), and is being hosted in this new format for the first time. Collocated with Ubicomp 2013 will be a conference covering the topic of “wearable computing”. Ubicomp 2013 will open with a speech by ETH Zurich Professor Markus Gross, director of Disney Research Zurich. The conference is sold out. No further registrations are possible anymore.