Is AI the solution to achieving the Development Goals on Sustainability?

The lack of progress being made to achieve the Sustainable Development Goals by 2030 points out that, without a magic trick up our sleeve, there seems to be no way we can meet those 17 goals – especially the 6 related to human influence on the environment. Our current track record says it all: 2019 was the second hottest year ever reported: from 2015 to 2019 the levels of carbon dioxide were 18 per cent higher than in the previous five years. 2019 brought us the coronavirus – a devastating zoonotic disease –  a clear consequence of the fact that Human activity has altered almost 75 percent of the earth’s surface, squeezing wildlife and nature into a small corner of the planet and breaking the barriers between human and wild animals who are the vectors of such diseases.

The only “silver lining” to all of this is that we have entered an era of unprecedented innovation and technological change, with massive development when it comes to Artificial Intelligence (AI) – a computer system that can be programmed to sense its environment, think, learn and act just as humans would. AI could be the “magic trick” we all have been waiting for, having the potential to provide the right tools for combating climate change, using ocean and marine resources wisely, managing forests, combating desertification, reversing land degradation, developing sustainable cities and providing clean affordable energy. To put it simply, achieve all the 6 environment-related Sustainable Development Goals. 

Climate change

Let’s start by looking at what AI can do to what regards climate change, the underlying and amplifying cause of most of the environmental problems known today. Some of the ways AI can help us treat climate change are 1) Optimising energy efficiency in buildings and decreasing the cost of clean energy; 2) Tracking carbon emissions; 3) Developing smart transportation systems and 4) Predicting fires. However, at this early stage, AI is not fully capable of solving everything. Most of the already developed applications focus on delivering insight for informed decision making, by analysing the unstructured real-time datasets available. 

In facility management, AI can help recycle heat within buildings, maximizing the efficiency of heating and cooling. An example of a company already providing this service is IBM’s Tririga.

Aiming at reducing the cost of clean power, AI is using machine learning to further develop “smart grids”- an intelligent electric system that monitors and optimises electric energy supply. Some companies already working on this are the ClimaCell, which employs machine learning to accurately predict weather patterns that can improve renewable power plants’ readiness and minimize safety risks, and DNV GL that uses machine learning analysis of data from sensors, attached to solar and wind power generation plants, to enable remote inspection of sites, prediction of maintenance, and energy resource forecasting. All of which end up reducing the final cost of renewable energy. 

Regarding tracking carbon, Carbon Tracker – an independent think-tank that is working toward the UN goal of preventing new coal plants from being built by 2020 – has partnered with Google, and is using AI to extend satellite imagery to monitor gas-powered plant’s emissions and get a better understanding on where air pollution is coming from, since the systems in place only monitor CO2 emissions near the power plants and not at global level.

For smart transport systems, we are already using machine learning algorithms that are employing car-sourced information to optimise navigation, such as Waze and Google Maps,  and to increase road safety,as Nexar, that provides car cameras that alert, record and share the accidents you have or will encounter on our path. In the long-run, AI will enable us to have guided autonomous vehicles (AVs), that will optimize routes to reduce driving miles and congestion, make use of eco-driving algorithms that prioritise energy efficiency and program “platooning” of cars to traffic. Platooning, to put it simply, is the capacity of cars to move in sync, accelerating and braking simultaneously. AVs, as you would imagine, presents a great opportunity to substantially reduce greenhouse gas emissions.

Fig: Truck platooning

Related to forest management, AI has the capacity to revolutionize the old ways of fighting fires. Through the years, we have experienced a drastic increase in the number of wildfires,  which are now burning more land than before and costing more to fight off. The traditional tools of spotting fires via planes or lookout towers, or simply having civilians report them, are considered highly inefficient, prompting us to more deadly disasters. One of the companies that is looking to solve this problem is Descartes Labs, based in New Mexico, US. They are using cloud base AI to analyse huge swaths of real-time satellite imagery to identify fires within minutes. They identify fires by spotting temperature anomalies, comparing the current temperature to the forecast of the region without a wildfire.

Fig: Figures on wildfires: number of them, number of acres they burned ,and cost to fight them each year

Healthy Oceans

As said before, all environmental problems are intertwined and stem mainly from climate change. Oceans for example are acting as a buffer for the adverse climate change effects, absorbing currently 20-30% of anthropogenic (i.e. human caused) carbon emissions and more than 90% of Earth’s warming. However, as you would imagine, all of this comes with a price. The price is that oceans are getting hotter. In fact, 2019 set new record highs for ocean heat, exceeding the previous record set in 2018. Moreover, the increasing levels of CO2 are changing oceans pH, making them more acidic – data says that oceans have become 26% more acidic since the beginning of the industrial revolution. The sad part is that climate change is just one part of the equation when it comes to degrading the ocean’s resources. We have not talked about overfishing, which creates an imbalance in the food web and pollution. Some figures point out plastics in the ocean can soon outweigh the fish in it if action is not taken soon. Obviously all the things mentioned translate into devastating consequences for marine biodiversity, and negatively impact the over three billion people who depend on marine and coastal biodiversity for their livelihoods.

Fig. Climate change effect on coral reefs

So saving our oceans must become a priority and we need to establish measures to reduce overfishing, marine pollution and ocean acidification. You may be wondering what AI can do to help us achieve these goals. AI applications are showing great progress when it comes to overfishing; an example is the company Global Fishing Watch, which is using data from vessel tracking systems that broadcast their marine position to monitor illegal fishing. They identify the vessels that manipulate their unique vessel number and GDP location and communicate their behavior to the marine authorities.

Also, to know the extent of our damage to the environment, companies such as Ocean Alliance, are programming drones to collect mucus samples for whales to obtain DNA information and measure the mammals’ health – thus extrapolating on the conditions of habitat in which they live. Some AI-powered robots are being developed that detect pollution levels and monitor temperature and pH changes. To efficiently clean plastic that is floating in the ocean, the non-profit organization Ocean Cleanup – you may know them because of The Interceptor that cleans river waste – has partnered with Microsoft to develop an AI solution that makes it possible to detect and quantify plastic in water via cameras on drones, ships and bridges. To develop this, they had to label around 25,000 images of ocean litter and use this data to create multiple models that could autonomously detect plastic pollution. They were already able to measure the scale of the plastic waste problem and now they estimate that within 5 years they could collect 50% of the ocean’s garbage.

Fig. Identification of ocean’s garbage

Biodiversity and land conservation

Human activities and climate change are accelerating both the deforestation and desertification of land. These are serious problems that are having major effects on our and other species’ livelihood on Earth. These effects include the fact that zoonotic diseases are more prominent now, and that from the known animal breeds 8% are already extinct and 22% are at risk of extinction in the near future. Nature is undeniably critical to our survival, and it is vital for us to invest in land restoration and ensure that we are using remaining resources sustainably. 

Even though we cannot go back in time and reverse the things we have done to the planet, we can now, at least, employ AI to make use of land resources wisely, detect land-use changes and prevent animal poaching. When it comes to using Earth’s resources wisely, AI has the power to make agricultural practices more sustainable by better monitoring and managing environmental conditions, and determining the best time to plant, harvest and use pesticides. Such technologies allow us to reduce both pesticides and water use, which is already a scarce resource. 

Fig: Intelligent agriculture

In terms of monitoring habitats and detecting land changes, AI provides an opportunity to detect the disturbances such as pests, damage, drought and fires. And lastly, AI can use machine learning to stream terabytes of videos and spot potential poachers and prevent them from killing animals. 

We cannot solve what we cannot measure, and this for now is the main advantage that AI (with some exceptions) is bringing to the table. In a few years AI will likely be programmed to autonomously make-decisions and not just analyse unstructured datasets, but this will as well come with more safety risks. To be sure that AI outputs can be trusted, governments and industry leaders must ensure that AI algorithms reasoning is transparent and that they can be explained to humans, who can audit those algorithm outputs and spot unintended biases. Also, AI needs to be well programmed to reduce the risks of hacking, that can be a serious threat for personal safety, especially if those hacks happen to autonomous vehicles. Achieving this requires a collaborative effort to ensure that AI progresses are aligned with human values and they encapsulate a future that is safe for humankind in all respects- our people and our planet. 


Alejandro. (2019, January 21). AI Is Helping Fight Wildfires Before They Start. Retrieved October 23, 2020, from Time website:

Choney, S. (2020, July 30). Microsoft Hackathon leads to AI and sustainability collaboration to rid plastic from rivers and the ocean. Retrieved October 26, 2020, from Microsoft website:

Gow, G. (2020, August 21). Environmental Sustainability And AI. Forbes. Retrieved from

Herweijer, C., & Waughray, D. (2018). Fourth Industrial Revolution for the Earth Harnessing Artificial Intelligence for the Earth. A report of PricewaterhouseCoopers (PwC).

Metz, R. (2019, December 5). Spotting wildfires is hard. AI could change that. Retrieved October 25, 2020, from CNN website:

United Nations. (2020). Forests, desertification and biodiversity. Retrieved October 26, 2020, from United Nations Sustainable Development Goals website:

United Nations. (2020). Climate Change. Retrieved October 26, 2020, from United Nations Sustainable Development website:

United Nations. (2020). Oceans. Retrieved October 26, 2020, from United Nations Sustainable Development website: