Agenda 2030 Graduate School blog

Lund University Agenda 2030 Graduate School is a global, cutting-edge research school and collaboration platform for issues related to societal challenges, sustainability and the 2030 Agenda. The 17 PhD students from all faculties at Lund University enrolled with the Agenda 2030 Graduate School relate their specific research topics to the Sustainable Development Goals. In this blog the PhD students of the Graduate School discuss topical research and societal issues related to the 2030 Agenda.

Can AI solve our environmental challenges?

Miniature tree among big trees. Photo montage.
Photo by Iván Tamás on Pixabay.

Posted on 20 May 2021 by Jesica López (Centre for Environmental and Climate Science) and Maria Takman (Department of Chemical Engineering).

The views expressed in this publication are those of the authors and do not necessarily represent those of the Agenda 2030 Graduate School or Lund University. The present document is being issued without formal editing.

This post is part of a blog post series on AI and sustainability.

“Building a more sustainable future requires a rethink of some deeply held assumptions about the role of artificial intelligence.”

Victor Galaz

In Dark Machines[1], Galaz explores why the idea of planetary change, digitalization, and AI, is naïve and dangerous. The hope that AI solutions can help “solve” deeply complex social, economic, and environmental challenges is described as just Hype. Our living planet is complex and it has its own capacity to sustain our growing needs, however there is a hyper focus on novel methods behind AI that ignores what is really important and it gets stuck in assuming a largely stable world. AI has potential to contribute to a more sustainable development in many fields, but AI alone is not always the answer.

In many industries, AI can entangle positive and negative contributions for environmental sustainability. Including agriculture, forestry, energy, transportation, water resource management, manufacturing, material science, and the list goes on. As the field of AI continues to evolve, so does its ability and applicability, appeal to help the environment. Just because the algorithms are all around us, as technology becomes more familiar daily. Indeed, it is possible to say that algorithms underpin almost all environmental monitoring technologies nowadays. In 2015, at the Transformations 2015 Conference in Stockholm[2],  our Planet Earth was presented as a geosphere that evolved to the biosphere and then, we humans created the Technosphere, as an integrated element of the biosphere. Moreover, at the same event the Biosphere Code Manifesto v1.0[3]represented a way to develop algorithms that helps us protect global ecosystems for all involved in the algorithm uprising.

Coming back to the “industries” and looking at the promising potentials of AI, for example in agriculture to have the power to make agricultural practices safer for the land and the health of the public or solving agricultural water efficiency problems[4] and the creation of and enabling environment for data-driven farming[5]. Moreover, another example related to monitoring of forests, where AI is helping local communities in using technology to protect the rainforest from illegal deforestation[6], and enabling diverse groups of stakeholders across the globe to have access to the latest data, technology and tools that empower them to better manage and protect forest landscapes[7]. In addition, the energy sector, where AI is increasingly used to manage the intermittency of renewable energy so that more can be incorporated into the grid[8]. Or, water resource management, where AI can help reduce or eliminate waste while lowering costs and lessening environmental impact[9]. Finally, if we talk about climate change, AI has been seen as a game changer. Through computer systems that “can sense their environment, think, learn, and act in response to what they sense” AI has helped researchers and governmental agencies to create weather forecasts achieving 89-90% accuracy in identifying, for example, tropical cyclones, weather fronts, and atmospheric rivers, all of which can cause heavy damages and are often hard for humans to identify on their own. It seems like AI is shaping itself to be a tremendously useful tool in our society. Although AI presents all these transformative opportunities to address the Earth’s environmental challenges, left unguided, it also has the capability to accelerate the environment’s degradation[10]

These environmental benefits for our society to get better climate forecasts, to make the jump into what is called precision agriculture/smart agriculture, the optimization of cooling of grids for energy, and even, but not forgotten, the monitoring of supply chains and global shipping[11], ocean mining, fishing, coral bleaching or the outbreak of a marine disease, is showing that AI has powerful advantages with some promising projects going on globally. Nevertheless, let us not set aside the so-called cons, better said risks, just because we are impressed by the tempting strategies and applications of AI in the environmental sector. Apart from the high cost of implementation, no human replication, no improvement with experience, the no-creativity nature of its algorithms, and maybe the riskiest of all with severe effects, the rise of capital-intensive technologies with less human-intensive requirements[12], AI’s potential to help address the climate challenge lies not in optimizing systems, but in augmenting people’s capacities to become stewards of the biosphere[13].

All being said, the pros and cons of AI in the environmental dimension being evaluated, we need to be more responsible together with the technology giants, international think-tanks and policy-makers to redirect and think ahead towards AI advanced agendas. As a good example, the potential of AI can be put in terms of how we handle nutrients and water sources. In fact, our relationship with water is close, a human body is up to 60% water and we live on a “watery” planet with about 71% of the Earth’s surface covered by water. We use it for drinking, for irrigation, for washing, for energy production, in industries, and probably in many other applications. One could perhaps even say that water is one of our most important resources. On the “opposite side”, or in “the other end” of the water we use, is the wastewater we produce, both our very own faeces and urine, as well as wastewater from a number of industrial applications.

Resources in water and wastewater

To protect the precious resource, which water is, and to protect ourselves from the wastewater we produce, different infrastructural systems have existed for ages. The oldest traces of systems to handle sanitation in communities are from the Mesopotamian empire (3500 – 2500 BC), and also the Romans are famous for their refined water and sewage infrastructure[14]. However, when the Roman empire collapsed, so did much of the water and wastewater systems. Pots were for example instead sometimes emptied directly out on the street, resulting in increased disease spreading, and the period from the middle ages to the industrial revolution is even sometimes called “the sanitary dark ages”[15]. With the industrial revolution and rapid urbanization, new sewer systems were once again built to decrease spread of disease and to improve sanitation in urban settlements. As environmental awareness increased during the 20th century, focus also shifted towards decreasing emissions of nutrients, and later also chemicals such as pharmaceutical residues and other micro pollutants. Consumption of energy and chemicals have also been put in the spotlight, and improved automatic control can optimize treatment processes and potentially decrease use of those resources.

In Les Miserables, Victor Hugo wrote that “the history of men is reflected in the history of sewers”[16][17]. Questions discussed at many levels in our society today, such as digitalization, AI, and environmental concern, are of course also reflected in the discussions about our water and wastewater systems. AI could be applied to save both money, energy and chemicals through optimizing different process steps in our water and wastewater treatment plants, and potentially enhance our understanding for specific processes, plants or waters[18]. There are initiatives to use AI to easier and faster track water quality parameters, such as bacteria and heavy metals[19], to use AI to optimize irrigation, based on moisture measurements in the ground and weather data[20]. AI is also tried out to assess status, maintenance needs, and risks of future malfunctions in piping networks[21].

The global challenges are however, unfortunately, sometimes more complicated than this, and there are of course also questions to which AI alone is not the answer. Challenges regarding safe drinking water and sanitation for all, resource efficiency, or for example land use are probably also on a more systematic level. To, for example, use the potential resources in the wastewater, such as water and nutrients, in a sustainable way, the challenges are probably in terms of both technology, politics, logistics and infrastructure, and the fundamental way of how we handle such resources might have to be changed, potentially including recycling of both nutrients (such as phosphorus and nitrogen)[22] and water[23], or other ways of more efficient use. Many wastewater systems are for example based on toilets flushed with water, in many cases of drinking water quality. In areas where drinking water is scarce, it is of course questioned if this is the best way of handling fresh water (even if you can use AI to optimize the treatment process or reduce leakages), or if other types of, less clean, water could be used, for example reused wastewater or sea water. Going back to the example about nutrient recycling, mining for the nutrient phosphorus may cause both environmental and economic problems[24][25], as well as being a non-renewable resource[26]. To reach a system change where we instead return our nutrients to the farmland, optimized irrigation or fertilization using AI will not alone be a solution, even if it probably has potential to help along the way.

In conclusion, many AI tools are developed “to solve” problems with land use, climate change, biodiversity loss, or water and wastewater management, but the bigger picture also needs to be considered. AI probably has big potential to optimize current and future water and wastewater systems, to optimize processes and decrease use of energy of chemicals, reduce operational problems, simplify water quality monitoring, optimize irrigation, and help in maintaining our piping systems. There are probably also challenges where AI alone is not the solution, such as how to handle important resources such as freshwater and nutrients.

[1] A book about AI, automation, and our living planet.

[2] Transformations2015, held in Stockholm Oct5-7 2015, focused on transformations towards sustainability.

[3] Galaz, V. 2015. A manifesto for algorithms in the environment.

[4] Itshaky, R. 2021. How AI will solve agriculture’s water efficiency problems

[5] Microsoft, 2015. FarmBeats: AI, Edge & IoT for agriculture.

[6] White, T. 2018. The fight against illegal deforestation with TensorFlow.

[7] Global Forest Watch.

[8] Polycarpou, L. 2013. Charting the Course to a Renewable Energy Future.

[9] Gow, G. 2020. Environmental Sustainability And AI.

[10] World Economic Forum. 2018. Harnessing Artificial Intelligence for the Earth.

[11] TRASE platform. Transparency for Sustainable Economies.

[12] Kumar, S. 2019. Advantages and Disadvantages of Artificial Intelligence
Advantages and Disadvantages of Artificial Intelligence | by sunil kumar | Towards Data Science

[13] Galaz. V. 2021. Will the Fourth Industrial Revolution Serve Sustainability?

[14] Lofrano, G., Brown, J., (2010) Wastewater management through the ages: A history of mankind. Science of the total environment. DOI: 10.1016/j.scitotenv.2010.07.062

[15] Lofrano, G., Brown, J., (2010) Wastewater management through the ages: A history of mankind. Science of the total environment. DOI: 10.1016/j.scitotenv.2010.07.062


[17] Lofrano, G., Brown, J., (2010) Wastewater management through the ages: A history of mankind. Science of the total environment. DOI: 10.1016/j.scitotenv.2010.07.062

[18] Zhao, L., Dai, T., Qiao, Z., Sun, P., Hao, J., Yang., (2020) Application of artificial intelligence to wastewater treatment: A bibliometric analysis and systematic review of technology, economy,management, and wastewater reuse. Process, safety and environmental protection

[19] Can AI help combat the world’s clean water crisis? (2019) ;

[20] Svensk startup ska vattna gräsmattor med AI (2019)

[21] AI vattenledningsnät (2020)

[22] Högstrand, S. (2020) Coming closer to closing the phosphorus cycle – how bacteria could be the solution to the coming food crisis. Journal of Water Management and Research.

[23] Takman, M. (2019) Opportunities for wastewater reuse in Sweden. Journal of Water Management and Research.

[24] Forskning från Marocko styr världens matproduktion (2019)

[25] Florida’s frightening phosphate problem – Phosphate mining’s significant threats to Florida’s water and wildlife

[26] Högstrand, S. (2020) Coming closer to closing the phosphorus cycle – how bacteria could be the solution to the coming food crisis. Journal of Water Management and Research.

May 20, 2021

This entry was posted in

AI and sustainability Okategoriserade