Prof. Kochupillai, a lot of your research has focussed on sustainable agriculture, while the work at AI4EO is concerned with Earth observation. What do these two fields have in common for you?

It’s the issue of ethics in artificial intelligence. At first, I investigated ways of promoting sustainable agriculture and the protection of agrobiodiversity, in other words biological diversity in agriculture. In doing so, I discovered that new technologies like blockchain and AI can play a part in this.

How did you then approach the ethical issues of AI in Earth observation?

At the beginning of the project I had a lot of conversations with the members of Prof. Zhus´s team and then also with researchers from other labs. I wanted to understand exactly what they do and where the challenges and limits to their approaches and methods are.

What did you learn from these conversations?

I’m glad to say that I encountered genuine interest in ethical issues. However, I also discovered that most colleagues, including some very experienced researchers, had no idea how or even why they should consider ethical issues – particularly those involved in basic research.

When it comes to ethical issues it makes a big difference if we think about basic research without a particular application or if we have a specific application in mind. There are different ethical issues and opportunities at the different stages of research, from basic to applied research. This is why we must think about all the issues throughout all the stages of research. The more concrete the application, the more specific the ethical issue. At the same time, ethical issues must be identified at an early stage of the research and addressed appropriately. At AI4EO the aim is to support various UN Sustainable Development Goals. So the work is no longer application-neutral but application-specific.

What specific questions do you derive from this?

Our questions refer to the type and source of the data in Earth observation, the labelling process, the method and location of data storage, but also access to the data and the analytical methods. In turn, each of these aspects can throw up further questions. Together with the team I am developing a schematic tool to enable us to understand the complex life cycle of the data employed in a typical AI4EO application.

In addition, we help researchers to evaluate the questions which underpin their work. We want to find out if possible unethical consequences or concerns which relate to the way the research is carried out can be addressed at the earliest stage of research. One specific question would be: What ethical issues and possibilities arise if we want to determine the function of a building or the degree of biological diversity in a farm field from space.

How do you go about this? What approach do you take?

We put the dominant ethical theories to the test. For example, in western philosophy, such theories are based on “obligation-based” (deontological) ethical approaches or on “consequence-based” (consequentialist) approaches. We research ethical theories from different parts of the world so that we can identify new kinds of approaches. 

It is difficult to apply the prevailing ethical approaches to novel technologies like the ones emerging at AI4EO. The researchers are not fully aware of their obligations and the potential future consequences of their research. In such a situation we must think about measures that we can introduce now in order to minimize negative ethical consequences and to maximize positive consequences. Admittedly this is a bit like looking into a crystal ball, because we don’t yet know exactly what we will be confronted with in the future. Therefore, we have to predict how things will develop and imagine what could happen.

What was the first outcome of your work?

I joined forces with four AI4EO researchers with different levels of experience to develop a tutorial about ethical questions and possibilities in AI4EO research. That paper contains a systematic explanation of ethics with many examples from current topics of AI4EO research.

Interdisciplinary research is an important element of the International Future Labs. How does this work in practice at AI4EO?

There is a challenge for AI4EO in this respect that is related to EU policy on agriculture. Under the “Green Deal” there is a major focus on sustainable agriculture and and agroecology, a system of agriculture that, by the way, is highly developed in India and is being applied systematically by a rapidly growing number farmers. This development over the past two years presents the AI4EO team with new challenges. For the sake of sustainable agriculture, the team needs to ask a different type of research question than it did before. Here we need a high degree of close interdisciplinary research between AI4EO and researchers working in agroecology and sustainable agriculture.

What obstacles do you regularly face in your research?

Currently, most AI4EO papers contain barely any content about ethics. Some researchers are afraid that inclusion of ethical aspects could cause their work to be rejected. This attitude needs to change! Specialist journals should demand research papers where ethics is a major focus and be proud to publish them. A change in culture is required in institutions whereby directors put the focus on ethics in order to ensure that ethics becomes integrated within AI4EO research. We really need a change of thinking in this.

What are the sources that you use for your work?

Ethics has its basis in philosophy, which is why I am constantly looking at philosophy from different parts of the world. But I also grew up with the traditional knowledge of India, with Ayurveda and Yoga. This is why I am always trying to combine the ancient and traditional with the new. My PhD thesis began with this quote from the ‘Ambassador of Peace’ Sri Sri Ravi Shankar: “Life is like a tree: Just like how a tree’s roots are old and the branches are new, in the same way, life needs ancient wisdom and modern scientific knowledge, both together.”

We, too, need a combination of ancient wisdom and new technology. We can combine them in research. In ethical terms, it would contribute to a better and more sustainable world. Ancient wisdom always includes consideration of sustainability, so that we take care of one another as well as the whole world. Around the world we can now see a growing interest in this traditional wisdom and sustainable growth.

Professor Kochupillai, thank you very much for talking to us. 

- The interview was conducted by Olga Gilbers -