COMEST workshop Ethics of IoT
“Our material world is changing rapidly. This influences us, and offers us new ways to ‘read’ our world. Technologies create a different enviroment.”
Introducing the topic is Peter-Paul Verbeek, professor of philosophy of technology at the University of Twente (and one of my PDEng supervisors). On March 14, Verbeek hosted a workshop jointly organized by the UNESCO World Commission on the Ethics of Science and Technology (COMEST) and the Department of Philosophy of the University of Twente. Goal of the workshop was to provide a basis for a COMEST report on the Ethics of IoT.
Verbeek: “Some ingredients of the day are sensors (think of perceptions, but also bias ) networks (provide us connections, and have vulnerabilities) data (can provide understanding, but come with privacy implications), and impact on society. Sensors help us make sense of the world. Using environmental technologies, we are creating a technotope (as opposed to a biotope). This forces us to rethink the relationship between citizens and governments.”
Paul Havinga: the technology of IoT
University of Twente, Pervasive systems
“I want to improve the world with technology. If you know what is happening, than you can do something about it.”
Understand → Predict → Act
This starts with monitoring. Use a wireless sensor network, with many small devices. The system as a whole needs to be reliable: it consists of a network of many parts, so individual devices don’t need to be as reliable; and the range can be small, because data can be sent through the network.
A big problem is energy usage and batteries. We don’t want more bits, but more bits per joule. Transmitting is costly in terms of energy, so transmit small packages. Processing costs less energy, so process locally in the network / in the devices.
Opportunistic sensing: A more recent development is the widespread use of smartphones. Everybody has them, and people keep their phones charged (so energy is less of a problem). The smartphone may provide lower quality data than dedicated devices. But you have access to many smartphones, making application cheaper and allowing much broader measures.
IoT is about: collaborative sensing, wireless and opportunistic networking and management, distributed data processing and reasoning.
An overview of IoT projects from the University of Twente:
Track the movement of cows using a wireless sensor network — how are the animals moving? is anything wrong? are they healthy?
Monitor the coral reef: a cheap buoy is equipped with a sensor-rope that goes down into the water. Many such bouys create a 3D image of things below the surface. The data quality of a buoy is much lower than a traditional data-logger. But it’s much cheaper, allowing for many buoys to be used, to produce a different view.
Using drones in firefighting to get a good picture of the fire: the drone has a camera and several sensors to measure smoke and air quality. The clothes of the firefighter are also equipped with sensors to provide more data-points.
Using underwater sensors to monitor oil rigs: if something goes wrong, there are large environmental impacts and economic impact. The sensors can monitor different things on and around the platform and underwater-construction. By creating smarter software, the battery could be halved in size.
Rob Kranenburg: IoT and society
European Research Cluster on IoT
“IoT is happening! It will never go away.”
Where are we in Europe? We are always keeping our eyes on the ball… we look and think, but we don’t act. The US is hitting always the ball, they have all the data lakes. On the other side, China is eating the ball — they take a very strong and complete approach. We have to find some realism: something between a transhumanist head-first all-in, and just saying “NO”.
Some figures for background:
1452 Gutenberg press. If the kings would have wanted it, we would have all been literate by 1500. Of course they didn’t want that.
1918 First public library in the Netherlands. It took over 450 years to distribute these learning tools.
1960s TCP/IP (built for speed): it cuts through everything.
1993 www: the first webpage is only 25 years old.
Existing systems and institutions are in decline. If you don’t own the data that you need for agency, you have no power. From the US, Alphabet (Google) is on the move to create a big cloud of everything. In 5 to 10 years, the nation state will no longer be relevant. China understands this, and is acting on it. China is running the counrty like a platform, building their own protocols and their own intranet.
What to do in Europe? The Next Generation Internet
Europe has to act, or become colony of either US or China. The notion of what power is, has changed. As long as we are trying to manage the old situation, we are becoming irrelevant. We have to really rethink what democracy is, what a state should do. We need a new set of protocols, we need a new type of United Nations.
The good example that we have is Estonia. After the wall came down, there was nothing. Three IT guys came back to the country, and built something new. We should now try to build something new before the current system comes down (in 5 to 10 years). Insert our own tools, and phase out the current internet and facebook.
Looking at China, they lack trust: they have cencorship within the data lakes, on the level of content — this is judgemental, and hinders innovation. What we (Europe) can do better is to look at what to centralise and what to decentralise. Data can stay with people — “I only expose my data when I want to”. All that needs to be centralised is a generic infrastructure. We have to build a middle space.
“Digital is a new ontology. We have to find tools and vocabulary, new notion of ethics, privicies, securities…”
Jeroen van den Hoven: IoT and Ethics
TU Delft, Ethics/Philosophy of Technology
“How to do ethics in the age of IoT, Big Data, and AI? Forget the funny gadgets, and focus on the urgent and important problem: the UN Sustainable Development Goals.”
Digital technologies are converging, and ethical issues are converging. We have to continuously look at the questions together. We have authoritarian regimes doing great, thinking in ‘central government’ terms. On the other hand, society is being designed by Engineer-CEOs. Europe’s counter is Responsible Innovation: try and solve a problem (using technology), without introducing new and bigger problems. Ethics becomes a question of design: Value Sensitive Design.
Moral characteristics of IoT:
- Pervasive (everywhere)
- Invisible (small, integrated)
- Ambiguous ontology (distinct natural and artificial objects)
- Identification (everything has an identifyable address)
- Connectivity on steroids
- Autonomous agency
- Embedded intelligence (extended mind)
- Seamless interaction and information flows (default is lots of data)
- Distributed control (networked, technology supports peer-to-peer)
- Complex & unpredictable: what is it “to act”?
- Big data (and privacy)
How to deal with this moral complexity?
Ruth Barcan Marcus: If you have a moral obligation to do both (a) and (b), you have a second-order moral obligation to see to it that you can do both. To innovate is a moral obligation: redesign your environment and your options.
A. gather information on the outcomes of your actions
B. Evaluate outcomes in terms of moral values
C. Use these considerations as design requirements for a new technology
We create technology — and in turn, the technology shapes us. If we do not design for our values, others will do it for us.
Nirvana Meratnia: IoT for animal protection
University of Twente, Pervasive systems
“IoT is NOT: network & connectivity; devices & things; chips & sensors; products & apps. It’s a paradigm shift: connectivity of everything, reasoning and making sense of everything, services, added value. It’s all of these things.”
- Living things → focus of this talk: animals
- Processes/business
- Things
- Data
Monitoring in Animal Ecology
- Eulerian approach: monitoring devices in the environment
- Langragian approach: monitor the animal itself (or herd of animals)
We want to protect animals, but how to do it? Some classical ethical approach run into different issues. Contractarian: animals don’t make contracts. Utilitarian: conflicting interests between the humans and the animals. Animal rights: leave all animals alone. Respect for nature: value of each individual. Contextual: what is the responsible thing to do and the value of things? Unclear how to continue…
Rhino case study: protect against poaching
Rhino poaching happens in different countries, and there are not many rhinos left worldwide. Rhino horns are used in traditional medicine, as trophy hunting, and the horns are now much more expensive than gold.
For anti-poaching there are obtrusive and non-obtrusive technologies, with several issues. Intruder detections and fence surveillance are expensive, and it’s unclear what you are detecting. Dye the horn or dehorning has impact on the animals. Use gps-camera implant, however, if the implant is hacked, the poachers can easily find the rhinos. Use drones, farmers shot down the drones.
Use other animals as sensors: monitor the behaviour of other animals. Measure and interpret disturbances in the environment and in the behaviour, and alert the park rangers.
Start at kinderboerderij: equip goats and sheep with sensors that can measure movement. Calibrate interpretation of data — what is walking, hopping, running? For IoT: the less data the better → get the right data! The question is when we can generalise and go from sheep to impala. Our hypothesis is that we can find overlap between the movement and behaviour of sheep and goats. After that include more other animals, until we can make a generic recognition model.
Two sides of a coin: IoT vs Ethics? Good and bad sides of IoT?
IoT can promote Ethics: use it to prevent poaching!
Marie-Hélène Parizeau: connected toys and children
Laval University Quebec, Philosophy department
“Connected dolls, toy, and robots are already here. This introduction of children in human-machine interaction is new in the discourse.”
Children regarded as vulnerable, therefor it’s relevant to look at ethical issues and responses, in the US and Europe (France and Germany).
In comparison to ‘normal’ toys and ealier ‘intelligent’ toys (e.g. Remote controlled car, limited and programmed interaction), connected toys can remain relevant longer, through updates and learning. Data-exchange is required, the toy can only interact if there is data gathered from the child (e.g. speach). So far, the focus of pediatric research has been on video-games and addiction. ‘Connected toys’ is an extension of the internet and data debate. 75% of teenagers have their own smartphone, 25% of them describe themselves as constantly connected.
US
Liberal society: legal and economic contractual approach. COPPA (protection of children under 13yrs) requires producers to delete data beyond immediate use. FBI warns against hacking of connected toys (e.g. spying barbie). Pediatric society advices to limit access to smart toys, depending on age and time used.
Europe
France: CNIL has a role to ensure compliance with law. They mention issues with relation to child’s private life and ‘secret garden’ (secrets that the child shared with a toy, not to be shared with parents). Spying by parents on the children can be seen as an invasion of the child’s privacy.
Germany: some toys have already been forbidden — they were compared to surveillance tools. However, security agencies have also asked/demanded to introduce a backdoor to have access to the connected toys.
Connected toys are included in the vast category of connected objects that keep pushing towards more human-machine interaction within ‘continuous connection’. ‘Continuous conection’ makes a ‘system’ within the meaning of Jaques Ellul, with its power, speed, effectiveness, and its invasion.
Connected toys are a good symbol for our society: To be connected has become a modern way of life that mediatizes more and more social and commercial relations but also private relations between people, parents and their children. This extension creates a form of continuum between public and private life.
Hagit Messer-Yaron: Opportunistic IoT
Tel Aviv University, Department of Electrical Engineering-Systems
“Opportunistic IoT: IoT which is based on opportunistic use of physical sensors and existing infrastructure. The new sensor can be a virtual sensor, such as our opportunistic rain sensor.”
Traditional rain measurement is done in three ways: with local gauges and weather stations — these are very accurate, but only give a local picture; radar — more expensive, monitors clouds, has limited coverage; satellites — has a good coverage, but is less accurate, and very expensive. In terms of global distribution, the US and Europe are well covered, but the rest of the world much less.
Idea: use cellphone signals to detect rain
When you make a call, your cellphone connects to a base station, the base station transmits to the next antenna in the network — until the last step, and is then connected to the other caller. The connection between base stations is called backhaul, and signals are transmitted in the microwave bandwidth. These particular signals are influenced by rain in a reliable manner. With more rain, the signal stregth deteriorates. Measuring the signal strength between two antennas creates a virtual sensor to detect rain between those antennas.
This virtual sensor gives a high resolution for a very low cost — to create a great alternative for traditional methods for detection and monitoring rain. Globally this has the potential to create over 4 million sensors, compared to ~400.000 weather stations worldwide.
Other than rain, the same method can be used to detect humidity, fog, dew, snow/sleet, pollution, birds (migration). Some applications for opportunistic IoT: meteorology, hydrology, reduce weather related hazards (e.g. aviation), emergency management, flood warning.
Opportunities and challenges
The biggest opportunities are sustainability (zero footprint), no need for special purpose equipment and big upfornt investments. In this case, there are also no privacy issues, because the signal used for measures is in no way tracable to the individual caller or the contents of the signal.
A big challenge lies in the discussion of public vs private in OIoT , and who owns the data. In this case, the phone company would argue that they should own the data, because the sensors are based on their infrastructure. Another perspective is to say that wireless bandwith is public space, the phone company only has the licence to operate in this bandwidths. In this line of resoning, the data created are a public good — related to the public space that the bandwidth represents.
Another case would be the use of smartphones — this represents private opportunistic IoT. Again this would pose questions related to the ownership of data, as well as future uses, consent, and transparency. Opportunistic IoT has inherent advantages. In the use of public resources, appropriate regulations can be applied. But the case points to the need to reset public-private discussions.
In Burkina Faso, there was a flood, with devastating effects. They had difficulty monitoring and preparing for such events because of lacking observational infrastructure. But they do have mobile phone antennas, allowing for this application of Opportunistic IoT.
Sevim Aktas: the Living Smart Campus project
University of Twente, DesignLab
“A Living Lab provides a space with more liberal regulations to test and evaluate innovations that are relevant for society, but that are also pushing legal borders.”
“Try out, to find out”
We live in a world of cyberbullying, lack of privacy, and impact on personal relationships. We accept facebooks terms and conditions without reading — facebook can do a lot of things with those data. Data is power and inequality is growing. Who could have guessed these things a few years ago?
the Living Smart Campus project provides an environment where innovations can be tested to evaluate consequence from actual application — instead of only from scenarios and thought experiments. The University of Twente combines a research environment, relaxed regulations, living & working environment, critical thinkers, and a safe-to-fail enviroment. At the core is transparency, in implementation, evalution, dissemination, replication, and refinement.
In order to involve policy-makers: involve and leverage the knowledge of social sciences students in every step of innovations. Leverage themes such as ‘smart cities’ to draw attention. Leverage the professors and management of the University of Twente, who are involved in the Design Lab.
What do we need to ask ourselves?
- Ethics during implementation
- Responsible innovation
- Experimental ethics
- Policy and regulations
