Things as Citizens in the inclusive smart city
The Internet of Things (IoT) has become a known concept for the smart city vision of the future. However, current implementations are often limited to sensing only, often in the form of the Internet as a dashboard, connected to sensors in the real world. In this arrangement, urban innovation stops at collecting data, visualising it via an app, and opening it for insights to base decisions on. How fast can the city adapt if things in the real world could respond to data? If they could not only sense but also act? What will be the role of things in the city? What are the ethics of things and can we create an inclusive smart city?
So many smart cities
Let’s first have a look at the smart city as a concept. The term is quite generic and there are a lot of different possible implementations of the smart city. With LABS we have been part of many types of smart city projects, from a social design driven project (The Civic City), to data driven research for smart shopping areas (De Nieuwe Winkelstad), and urban focused adaptive interventions (Co-ReUS). The relation of the city with its user and vice-versa is of course often the key factor. This can be seen in the projects on creating responsive data layers to help people exercise more (BAMBEA, PAUL).
The smart city can be classified by the type of defining element and while important, data is not always the center of smart city solutions. When data is the leading principle, it is seen in smart cities with open data platforms and data studios. Another principle is human relations, as cities are built on the people that live in them. In smart cities these relations become more intelligent and connected, and when done right, should result in more diversity. The social fabric of a city is also connected to a digital equivalent which is demonstrated in sharing services. Bike sharing plans or Peerby sharing service for neighbours are examples. The fabric of the city, its fixed and moving objects, encompass another defining element. Objects can function as hubs for people to connect or as artefacts for building relations. In this way, the objects can take on a new function as a platform for services.
Another interesting angle is the perceived goal of making the city smart. As Adam Greenfield it puts in his latest essay on the Internet of Things, Rise of the Machines, the typical ambition of the smart city is ‘control’
“If the ambition beneath the instrumentation of the body is ostensible self-mastery, and that of the home is convenience, the ambition at the heart of the smart city is nothing other than control — the desire to achieve a more efficient use of space, energy and other resources.”
This is true from a governing perspective, but I think we can add some other ambitions:
The Efficient City: This can be to create an efficient city with less congestion.
The Safe City: Strive for a more safe city (perceived), with focused surveillance, optimised lighting, and insights into citizens’ behaviour.
The Social City: Another approach is trying to create a more social or creative city by connecting it’s citizens or by using interactions between citizens to create a more pleasant neighbourhood.
The Healthy city: The healthier city, or better said, citizen, is a strong driver. This is seen when the smart cities gives insights and triggers for healthier behaviour for the people or uses insights to clean the air.
Towards an intelligent city
So in the current smart city vision, the data layer is seen as part of the city, a sensing city with dashboards of data. In the next iteration a new urban layer is added to the DNA of the city: the intelligence of things. As objects of everyday use become more intelligent and adaptive (from urban furniture to self-driving vehicles), things will become part of the city infrastructure and begin to work (hopefully) in concert with people.
The relation of the things to the data it collects can change. More and more insights are gathered from the sensors in the city. Sensors that are connected to the fixed city structure measuring ‘analogue objects’ or sensors connected to the things in the city measuring the movements and behaviour. The current smart city is in that sense a reactive system, delivering data for decision makers or smart systems to adjust the city scape.
In the model of Products of Agents, there are three levels of interactions between things and their context/users. Collectors are things that gather and present data, Actors have rules in them to use the data for the functioning. Third category is the Creators. They come up with new functionality based on their learnings and knowledge. Most smart city concepts are based on the Collector-model, sometimes it feeds the data directly into the behaviour like lighting that is adapting to the persons walking by, and the city becomes an actor on itself (or the system of objects at least).
An example is a smart lighting concept in Amsterdamse Poort shopping area. On a square all light posts got sensors to measure crowds moving on the square. Combined with data sources the light posts adjust the level of lighting and place where they shine following a couple of algorithms. The aim is specifically to create a safer environment.
Next level is the smart city as a Creator. Or better said, the things in the city as creators. We think this next step will be enabled through more intelligence that is integrated in new products, objects, and also the city things. These things will have agency on their behaviour, not only following one dimensional rules. Machine learning becomes deep learning. At the moment this happens we will see some challenges arise.
_ How do we interact with these things, and how do we live in concert with the things?
_ What is the frame of reference of these things? Can things be good citizens?
This is the premise of a research program we are setting up at Delft University of Technology Connected Everyday Lab. We will dive in this phenomenon and focus on different angles both from the thing perspective, the citizen perspective and also the cityscape infrastructure.
The end goal is to give the things an active role in purifying the air of the city.
Things as citizens
What happens as the things become more self-aware, more intelligent, have more agency to decide, and become so-called objects with intent? The things can become citizens in the city just like us humans. The relationship we have with things will change under the influence of the intelligence they can embed and use. While artificial intelligence (AI) has been researched for many years, recently ethics in AI are increasingly important as digital services are becoming more defining in our lives. The thing becoming an intelligent thing raises comparable issues.
Nick Bostrom and Eliezer Yudkowsky describe in a paper on the ethics of AI the principle of Moral Status of machines. We relate to things in several levels, and they define a 1–7 scale of moral status. Level 1 has no moral status, like a rock, and we can easily treat this rock in all kind of ways. Level 7 on the scale is our human moral status. Currently AI has no moral status, but that will change. The influences of moral status are two-fold:
_ Sentience: the capacity for phenomenal experience or qualia, such as the capacity to feel pain and suffer
_ Sapience: a set of capacities associated with higher intelligence, such as self‐awareness and being a reason‐responsive agent
It is not only about the way we approach the things with AI. There is a question of duplexity. The moral status is an indication how we may treat the AI. It also goes the other way; how does the AI treat us, and to what extent do we influence it.
This is where we can see citizenship-concept appear. A thing with intelligence and therefor a higher level of moral status needs to live up to that position by contributing to the city as a whole.
Ethics of things
In the Internet of Things community is a lot of discussion on the ethics of the connected things and the dangers it unlocks when everything is connected and collecting data. With ThingsCon we try to emphasise the discussion on a responsible and human-centric Internet of Things, and in London the Open IoT definition is developing a IoT certification to stimulate good design and manufacturing of things and the services that are part. The afterlife of a product is now integrated part of the responsibility of the brand. Just Things organisation made an IoT Manifesto on ten principles to take into account for designing things, that is recognised as a good point of departure for this thinking. It does not cater the intelligent things with agency completely.
All of these ethics discussions deal with the perspective of companies and brands that market and exploit the things and use the new possibilities of data collecting as a source for insights and possible bad motivations.
One aspect is often not taken into account; the ethics of the things themselves when operating in autonomy. We can easily transpose the intentions of the manifesto to this situation, on the other hand these intelligent things are maybe more comparable with humans and we should look at the moral compass we expect from humans living in the society.
An interesting issue arises as these moral compasses are often culturally defined. We have different values on privacy for instance in Germany and England, countries that share a lot from Western culture. Experiencing the culture in China makes you realise how extreme the edge cases can be. Sharing data by companies like Tencent (maker of abundantly used Wechat app) with governments for surveillance is perceived totally different from most European countries. Things will cross borders though, and will have to adapt to these cultures. Not as a touristic habit, but to really understand and embody the cultural paradigms.
How to design the AI is an important factor here. AI is defined by the aspect that it is essentially a learning system. This means that we don’t define behaviour and presence in the details, but we set the rules for AIs to evolve on. Rules to grow and boundaries to stay within. The learning process is defined by the interactions with other things and humans.
The discussions on liability of self-driving cars is one of the interesting shaping elements; who is responsible for the behaviour of an autonomous operating object?
Human — nonhuman interactions
As we look to the things as autonomous objects, or as things with agency, you can look at the interactions with humans as a defining element.
What differentiates agency from assistance is the autonomy and authority to act on a users behalf, without the need for user input. (source)
That is true, and it is also a good indication of the boundaries. In this model the human is the boundary, having control of the thing, and actively delegating authority is its concept. What happens if the thing is evolving behind this delegation?
Facebook had an interesting experience here last week when it turned out that the AI they were developing created a new language that only was understandable by the AI itself. Facebook decided to kill the project for now. (source)
It is an important part of rule-based design thinking. A rule-based approach can deliver a more natural feel in cities proposed in this article. That is interesting; not per se new for urban design as this is often based on storytelling translated in rules for architects. The power of rules as the system of living is nice and underlining the way AI is shaping reality or could, in dialogue with people. The concept of designing curiosity as a ruling principle is something to think about. As we focus on creating things that have an ‘open mind’, are curious to adapt, and are really interested in the values of the other, it could generate some interesting design approaches.
Kars Alfrink used this approach in fighting fake news; it is therefore interesting to compare this with preventing the things that will be built upon algorithms becoming evil.
A specific element here is the relationship of humans and things (or ‘non-humans’). With the rise of intelligence, the dialogue will become a dominating interaction principle.
The design of Objects with Intent must be primarily concerned with how tasks and judgments are delegated between humans and non- humans, and how this delegation is expressed in sensible form. (Marco Rozendaal, 2016)
We need to understand the reasons of the behaviour of things as they operate more on their own. It is not only about good behaviour, but it starts by understanding. What choices are made by the things, and why. I think it will be one of the principles in all design for AI; open up the black-box for the users so they can understand the decisions made by the AI.
Towards an inclusive city
So concluding we see that the role of things in the city is changing. In a broader perspective is the development towards a smart city not per se leading to a fairer city. The divide of people living in cities can become bigger. The luxury of having agency in your own life is often seen as a future we will move into. The haves and have-nots are not separated on technology or wealth in the end, the possibility to be yourself and to rule your own life will be more and more dependant on the possibility to disconnect from all sensors that surround you. Social Cooling is a nice framing of this new state of freedom.
Just like humans the non-humans in the city will have this separation. Cheap products that have weak security play a role in breaking the freedom of its users. The things with responsibility will be the good citizens that life up to certain standards and can elevate the people in the city.
As the dangers of the smart city include segregation and new class societies, it is necessary to introduce the concept of the inclusive city. A city that actively is involving all groups of citizens to create a kind of harmony.
That means we need to think on ways to include weaker groups, create a space in the city for all, not only the happy few. This is part of the strength of the city, the diversity and it triggers the right creativity as it is part of a positive collaboration.
It is a valid strategy to include the things in this mix. Things take a role in this city fabric and can become proxies for the inclusive behaviour a city as a whole can have towards its inhabitants. The challenge now, is to create a set of design rules for urban and industrial designers to reach this inclusive mode.
