How can AI help us tackle climate change?

Global efforts to combat climate change have been stifled recently, with an unprecedented crisis shifting the focus of governments, businesses and consumers away from the plight of the planet. We’re engaged in a war against an invisible enemy; and COVID-19’s dire importance has placed many environmental initiatives on hold.

Some may argue that the lockdown, and ensuing drop in carbon emissions, itself represents marked progress in the fight against climate change. However, this is only a temporary victory – once this virus is contained, carbon emissions will return to normal levels. As such, even during the pandemic, plans must be made to ensure a more sustainable and responsible battle against climate change is fought.

I would argue that the greatest resource we possess when developing such plans is artificial intelligence (AI). Not only does it provide the digital tools we need to effectively monitor and predict events, it also gives us the ability to design solutions that can effectively reduce our collective carbon footprint.

To illustrate this, below I’ve listed several examples of ways in which AI, together with machine learning (ML), is already assisting humanity in the struggle against climate change.

Known unknowns

An often-overlooked feature of AI comes purely as a result of its artificiality. An AI can never be accused of wishing to advance one particular agenda over another or from exhibiting an unconscious bias – both issues which are commonplace in the debate surrounding climate change.

This makes it incredibly useful when choosing between multiple valid models when tough decisions must be made. A group named Climate Informatics regularly discusses how we can best use the technology in this way, and their founders have long been exploring how to combine the approximately 30 different models used by the UN Intergovernmental Panel on Climate Change to produce one globally-agreed-upon statistical model.

Such a synthesized, definitive model would greatly aid intra-governmental efforts in standardising their measurements and informing policy makers through clearer international comparisons.

Whatever the weather

The most immediate way most of us will encounter the effects of climate change will be through extreme weather conditions. Events like floods, intense storms, or terrifyingly strong winds will become all too regular over the next century – so it behoves us to prepare for it as best as we can.

For decades, data scientists have been gathering data from our land, oceans and atmosphere to forecast future events and how they might impact our ecosystem. Yet the sheer volume of data required has previously rendered it extremely difficult to make useful, accurate predictions.

The rise of AI, however, means that we now have a powerful new tool at our disposal. Trained algorithms can sift through huge swathes of data and unearth the valuable insights held within. Importantly, they can do this at speeds unmatched by humans – and with much greater accuracy.

For example, researchers from Microsoft, the Montreal Institute for Learning Algorithms and ConscientAI have used AI to run simulations of how cities will be impacted by various degrees of extreme weather – right down to individual homes. Such exercises can provide incredible insights for city planners, policy makers and emergency services to help best prepare for, or even prevent, such catastrophes. They also provide the added benefit of illustrating the real dangers of climate change to the general public – hopefully to the extent that the threats it portrays may never come to pass.

Measuring the harm

Evaluating the harm caused by greenhouse gas emissions is half the battle in combating climate change. Increasingly, researchers are relying on AI technology to help gather in-depth data on the emissions being pumped into the atmosphere from the world’s power plants.

Some people not acquainted with AI often fear that the more advanced a system is, the opaquer it may be. However, the independent think tank Carbon Tracker demonstrates the exact opposite with its monitoring of coal and gas-powered plants using sophisticated AI and ML techniques.

Through processing satellite imagery, AI-powered algorithms are used to spot and quantify emissions from power plants across the world – including those that would generally fly beneath the radar. These algorithms are trained to measure how much pollution is being emitted, with researchers and activists able to monitor areas in real-time that they would never have had access to before.

Without accurate information, it is difficult to hold industry players to account and ensure they are adhering to their emission reduction targets. Applying ML to satellite imagery has transformed the way we are able to detect signs of power plant emissions and locate the worst offenders.

A recently gifted grant from Google will undoubtedly aid the think tank’s efforts, as well as its intention to share all of its information with the public – ensuring AI will be used to increase corporate transparency over carbon emissions across the world.

The examples listed above are only the tip of the iceberg when it comes to the many ways that intelligent technologies can be used in the climate crisis. AI has already facilitated gigantic leaps in sustainability technology, so there is no reason to doubt that it will continue to do so both during and after lockdown.

The drastic long-term changes needed to begin repairing humanity’s damage to our planet may seem insurmountable to many, but the greater the AI tools we have at our disposal, the more manageable this target becomes.