Artificial intelligence for sustainability: “Technology is the easy part”

Artificial intelligence has a lot to offer for sustainability. While the technology itself is complex, the real challenge is to highlight the benefits and create common rules to make this shift happen, claims Laura Ruotsalainen from the University of Helsinki.

According to the United Nations, sustainability is the number one priority for the international community. For science, this means developing solutions that enable a good life for our planet and people today, without compromising the possibilities for future generations.

One of the challenges of moving towards a sustainable world is that these solutions must be based on a great amount of diverse data from different sources.

“Our capacity to handle this data is limited: this is where artificial intelligence steps in”, says Laura Ruotsalainen, Associate professor at the Department of Computer Science at the University of Helsinki, HELSUS and FCAI.

“AI is capable of finding connections and tendencies that people would not even look for.”

Social sustainability requires un-biased data

Utilizing AI to create new solutions may cause some concern about the quality of the data. First, we must guarantee the reliability and privacy of the data, as well as access to it, and common rules for sharing it. For social sustainability, it is also essential to avoid bias.

“The data must cover all the relative points of view, also from different minorities – not just one version of our reality”, Ruotsalainen says.

However, solutions developed by people may also be biased, for example to fit our values. With AI, we can reduce this risk by assuring the quality of data.”

What about people and business?

From the technical point of view, some of the issues related to sustainability could be solved quite soon. For example, there are already smart mobility solutions that could substitute private cars. But are we willing to use these solutions?

“To make the most of them, the end-users have to see their benefits. The solutions must be user-friendly and ethically acceptable from all aspects of sustainability”, Ruotsalainen says.

The same goes for business.

“Data does not wear out like other commodities. But if a company has built its business around data, it should be able to enjoy the benefits. We need fair business models to encourage companies to create these solutions.”

From the ecological point of view, AI can consume a lot of energy for data processing.

“We must keep this in mind when we develop new AI methods. We should not aim for constant technical improvement if this increases energy consumption. Negative environmental effects should not outweigh the benefits.”

A bridge between research and industry

Fundamental research on AI is a pre-requisite for new innovations based on AI methods. For sustainable solutions, this often means combining different fields of research. For example, Ruotsalainen is currently working on a project called CousCous, which combines research on socio-economic factors, air quality, and AI methods. The goal is to simulate, model and optimize dynamic routing for sustainable mobility.

On top of this, we need collaboration between research and industry, Ruotsalainen says.

“Finland wants to be a leader in the application of AI, and we have good expertise both in research and business. But we need public funding to enable and enhance collaboration and build a bridge between research and industry. Companies cannot be responsible for this.”

The technology behind new sustainable solutions is very complex itself.

“But when you compare it with the whole system and the challenges of bringing these solutions to our society, technology is the easy part.”