How can AI help tackle climate change?

AI and machine learning is an enabling technology that has a critical role to play in reducing the carbon footprint of energy generation, transportation, food production, industrial manufacturing, and our homes, offices and cities. But is it enough? And can technology alone save us?

A resident cries as the Thomas Fire approaches the town of La Conchita, CA in Dec 2017. (https://bit.ly/37WD4K4/)

The EU parliament declares a climate emergency

On Nov 28th 2019, the EU parliament declared a global climate and environmental emergency. They say that all politics is local and across the world climate change seems to be coming home to roost. In the hills around San Francisco the bankrupt PG&E power company pre-emptively shutoff power to homes for several days as it worried that its ageing electrical equipment would act as a match to the parched trees and vegetation. In Europe extreme flooding has been immersing ancient towns in apocalyptic scenes. In Australia it was hard to discern the iconic Sydney Opera House for all the smoke from the raging bush fires. And in New Delhi, India officials declared a public health emergency and distributed five million masks to deal with toxic air pollution. And this has all happened in just the past few months.

It is becoming harder to ignore the issues of climate change. But what role does technology and AI play in addressing this?

A portfolio approach to climate change technology and AI

We know that we need a portfolio approach to technology development and deployment to address climate change. This portfolio can be dimensionalised across time, risk and maturity of technology. In the near-term the world needs to continue its push to develop and install renewable energy sources such as solar and wind energy power generation. In the mid-term we need to make bigger breakthroughs in the development of higher capacity batteries to store energy and power electric vehicles. Autonomous vehicles and smart cities can make a major difference to a reduction in carbon emissions but the technology needs to mature before it will be trusted for use in our daily lives. In the longer term we need fundamental research into truly transformational energy sources such as fission nuclear reactors. Societies need to be thinking of climate technology “moonshot” projects that will require significant financial investments but with the potential to address the scale of the issues. And across this portfolio AI will be a key enabler to all of these technologies.

Five territories for technological innovations

As the WEF discussed in 2017 it is critical to focus technology development on five areas of our society to address climate change:

1. Power and electricity generation account for 25% of all greenhouse emissions. A sustained and increased focus on modern renewables such as wind and solar power is critical. The WEF graph below shows that while modern renewables (in green) have shown strong growth since 1974 we still have a long way to go.

There is much research and development to be done in renewables from improving the cost and effectiveness of wind and solar to looking at marine, geo-thermals and biofuel energy sources. There are opportunities to expand nuclear power and also continue to focus on fusion technology moonshots. However issues of high costs and concerns over safety have slowed nuclear power deployment.

2. Transportation represents 23% of global related carbon dioxide emissions. The public is fascinated with the benefits of electric cars, buses and trucks, but there are two major problems that need to be addressed to bring these vehicles to the masses — (1) reduce the price of electric vehicles and (2) increase the capacity and reduce the charging time of batteries. There is also a focus on alternate fuels such as biofuels with companies like Virgin Atlantic airways claiming the world’s first waste-based biofuel commercial flight in 2018.

The Beyond Meat Burger

3. Food and its supply chain represent 25% of global emissions. This is not surprising when there are seven billion people to feed on this planet. Agriculture has seen swathes of forests razed to support cattle grazing and grow cattle feed that is sprayed with chemicals to optimise yields. Technology investments have included a focus on meat substitutes that use advanced plant protein science to replicate that unique umami taste generated from complex chemical reactions when meat is cooked. And some are including beetroots — beets — to make the burgers “bleed.” The Impossible Burger and Beyond Meat are two new companies that use plants to achieve a meat like taste — try them, they’re not bad but still quite pricey and not quite as healthy as we might imagine. And according to a recent University of Michigan report, a plant-based burger generates a whopping 90% less greenhouse emissions, 46% less energy, has 99% less impact on water scarcity and 93% less impact impact on land use. The next frontier is lab-grown meat where beef, chicken or salmon proteins are grown in a petri dish. All the while there is a focus on technology to improve the productivity of existing agriculture practices.

4. Manufacturing in industry represent 30% of global related carbon dioxide emissions. The world keeps adding factories to produce the consumer and industrial goods to sate the appetites of an increasingly wealthy and middle class planet. Technologies are being developed to help manufacturers become more productive and reduce their use of power and their toxic emissions. Carbon capture technologies, in particular, have been touted as a way to capture emissions. There is still a long way to go.

5. Buildings and cities represent 20% of global emissions. Think of all the lighting, power, heating and cooling used by our homes and offices. How do we switch to more efficient and even emission free heating and cooling systems? Google has played a role here with smart thermometers for the home such as Nest that learns our energy usage habits. And smart city technologies are being developed. The Chinese Hanghzhou ‘City Brain’ project is focused on reducing traffic emissions by optimising real-time traffic flows. Google’s Sidewalk Labs in Toronto claims to be reimagining the whole city.

Ten examples of AI being used to address climate change

AI is a key enabler of climate change technologies. While AI is not well understood and the media narrative often focuses on the negative consequences of job loss and ethical biases, AI is going to be woven into the fabric of climate change technologies.

The latest wave of AI and machine learning focuses on applying algorithms to masses of data to make systems smarter and higher performing . In its latest incarnation AI — machine learning and deep neural networks — is driving:

• Measurement and predictions on steroids — while data science and statistical techniques have been used widely in climate science to help optimise, analyse, assess, classify and predict, machine learning puts this into a different league. The data that is being generated by sensors in industrial plants, supply chains, human wearables and mobile phones all has a critical role to play in collecting information about climate change. This data can be then fed into more sophisticated models to help us better understand, target and manage our climate actions.
• Cognitive superpowers such as seeing, hearing, reading and understanding. In particular, the “vision superpower” can be applied to all sorts of important climate change activities such as viewing and analysing remote sensing data to identify pollution levels.

Ten examples of AI use cases that are working or are being worked on today:

1. Build better client models. AI is being used by the likes of the US National Oceanic and Atmospheric Administration to build better predictions of extreme weather events such as hurricanes and unlock new insights from the myriads of climate data that is collected.
2. Increase monitoring, measuring and accountability of pollutants. AI is being used to automatically analyse data from IoT sensors and remote sensing data to identify pollutants such as ground-level ozone, particle pollution, carbon monoxide, sulphur dioxide and nitrogen dioxide. This is particularly important in more remote locations where it is hard to determine the volume and source of pollutants. It will provide not only a more accurate and transparent picture of pollution but it will help drive accountability.
3. Optimise traffic flow and make smarter cities. AI is being used to measure and optimise traffic flows in cities. Just reducing the number of cars stopping at red lights by optimising flows through better timing of traffic signals can have material impact on carbon emissions. For example the, Chinese “Uber”, Didi is tackling traffic congestion and optimising navigation routes with deep learning. In India McKinsey & Co reports that smart rubbish bins are being trialed that means that rubbish trucks do not pick up if the bins are not 75% full.
4. Improve building energy consumption with intelligence AI powered devices that measure, predict and control heating and cooling systems based on actual need. For example, St. Vincent’s Hospital in the US achieves 20% in energy savings by implementing a predictive energy control system for its heating and air conditioning system. Many cite the example of DeepMind using reinforcement learning to reduce the power consumption of its parent, Google’s data centres by over 15%. Combined with smart AI powered meters and appliances we should see large reductions in power usage in our buildings.
5. Rollout autonomous trucks — while most of the discussion on autonomous vehicles has focused on autonomous cars, the most likely near term use will be in autonomous trucks. It has been estimated that smart automated driving systems could see a 15% reduction in fuel consumption over human operators. Similarly regulators are more likely to approve autonomous driving on major highways as they are more predictable and safer than the chaos and unpredictability of driving congested city streets.
6. Better match electricity supply and demand in a smart grid that allows communication between networks of consumers, transmission lines, substations, transformers and suppliers. AI will be key to better predicting and controlling supply and demand across such a complex network. For example Google’s DeepMind developed a deep neural network system that increases the value of wind power by 20% by predicting supply based on weather forecasts and historic turbine data. Their model recommends how to make optimal hourly energy delivery commitments to the power grid a full day in advance.
7. Improve efficiency of logistics and supply chain by better matching supply and demand. AI is increasingly being used to understand demand levels across complex and increasingly international supply chains. As an example, Otto — a German ecommerce company — predicts with 90% accuracy what products will be sold within 30 days driving automated purchasing and reduction of annual returns by over two million packages. In another example Schneider Electric reduced transportation needs significantly with an AI model to predict the best way to acquire raw materials and send their products to 240 global manufacturing facilities and 110 distribution centres.
8. Optimise the food supply chain and improve agricultural yields. This is an area that AI is already having an impact. From better predicting demand in restaurants to reducing food waste to helping developing world farmers diagnose and treat agricultural crops, AI is getting to work. For example, Microsoft has worked with Indian farmers to achieve up to 30% higher yields with machine learning advice on when is the best time to sow crops.
9. Improve manufacturing efficiency by digitising, connecting and analysing end to end manufacturing processes. For example many global manufactures are using predictive AI modelling to make turbine combustion more efficient, reduce errors and energy wastage on the production line, and improve production efficiency with advanced robotics.
10. Help consumers reduce their carbon footprint — AI driven consumer applications are being offering to help us measure and predict the level of our carbon footprint. Using gaming mechanics we can start to compare our sustainability footprint compared to others. This was brought to life, rather frighteningly, in a recent short-film from Ann-Catherine Beyer who imagined a world where we all gain or lose “Environment Credit Points” based on our behaviour and its impact on sustainability.

Ann-Cathrine Beyer, member of “Econtrol” film team; winner of Young Talent Award at 2019 “KI Science Film Festival.” http://www.zak.kit.edu/6427.php

For all the good of AI and machine learning we also need to be aware that computers use a lot of electricity. Some argue that information and communication technologies contribute up to 8% of global energy use.

Technology alone will not be enough

We know that solutions to systemic problems such as climate change are not simple. It will also require an urgent and integrated focus on:

1. Consumption patterns — the world’s population is expected to increase from seven to ten billion by 2050. It is critical that awareness is raised of our own individual carbon footprint through our daily choices around what we eat, how we travel and how we live. But we also need to recognise personal responsibility is set in the context of great sections of society who are escaping historical scarcity and entering an aspirational middle class lifestyle. Look to China where the middle class is now larger than the whole US and citizens are consuming with, what some might describe, as wilful abandon.
2. Multinational coalition and regulatory policies — the world’s governments and organisations need to work together to set clear goals and policies such as the 2015 Paris climate accord. But a lot more will be required. We will need much further legislation around pollution, consumption and energy usage that no doubt will be viewed as impinging on personal choices, liberties and free-markets. This legislation will be hard to do in an increasingly polarised society where wealth inequality is driving a rise in populism against establishment decrees.
3. Financial incentives — the cost of dealing with climate change is in the trillions. Even though China invested $100B in clean energy in 2018 and the US $64B this is a fraction of what is needed. And with a financial system generally focused on short-term results, such as quarterly growth and profits, the question becomes how do we realistically incentivise governments, companies and organisations to invest in the longer-term. The increasing momentum behind impact investing is helping to bring capital flows to those organisations that are linking environmental, social, and governance” (ESG) goals with financial returns. At the multi-lateral level Christine Lagarde, the new president of the European Central Bank, is spearheading a global drive to make the environment an essential part of monetary policymaking. And the World Economic Forum (WEF) in advance of Davos 2020 just called for a “better type of capitalism.” They are advocating “stakeholder capitalism” with new measures of “shared value creation” that include ESG goals as a complement to standard financial metrics. There is momentum. The US Business Roundtable, America’s most influential business lobby group, surprisingly and admirably, called for a form of capitalism that moves beyond simple measures of profitability earlier this year.

AI will be a major enabler of technological innovation to address climate change along with other fourth industrial revolution technologies such as smart materials, autonomous vehicles and internet of things. It will enable more efficient power generation, smarter cities and buildings, zero carbon transportation, enhanced food supply chains, and more efficient and carbon neutral manufacturing. However technology is not enough in itself. We need to change our individual consumer consumption patterns and not rely on simply assuaging our guilt through the purchase of carbon offsets. The political and financial systems will be challenged but we need to be pragmatic and know that human and corporate desires and behaviour will be slow to change — we want our bigger houses and companies want their profits. Here’s hoping that we can act fast enough.

Thanks to recent panel discussion at the University of Sussex

This article was inspired by a recent panel discussion I joined at the University of Sussex, England, on the question of impact investing, technology and climate change on Nov 29th 2019. Panelists included Neha Coulon, Head of EMEA Capital Strategies in the Sustainable Finance division at J.P. Morgan, Prof. Jeremy Hall, Director of the Science Policy Research Unit (SPRU) at Sussex University, Dr Matthew McCarten, Spatial Finance lead at the University of Oxford, Jessica van Thiel, Founding & Managing Partner of Pathfinder, a social enterprise for sustainable development solutions, and Alex Martial, a third-year undergraduate student at the University of Sussex Business School. A big thanks to the moderator of this panel Isabel Fischer.

Useful AI, technology and climate change articles

I found the following articles and resources useful:

1. McKinsey & Co on how technology is driving new environment solutions
2. World Economic Forum on 5 tech innovations that could save us from climate change
3. Bill Gates on this is what we need to do to tackle climate change
4. David Rolnick et al on tackling climate change with machine learning
   (see below)

About Simon Greenman

Simon Greenman is a partner at Best Practice AI — an AI Management Consultancy that helps companies create competitive advantage with AI. Simon is on the World Economic Forum’s Global AI Council; an AI Expert in Residence at Seedcamp; and Co-Chairs the Harvard Business School Alumni Angels of London. He has twenty years of leadership of digital transformations across Europe and the US. He holds a degree in Computing & AI. Please get in touch by emailing him directly or find him on LinkedIn or Twitter or follow him on Medium.

#sustainability #environment #climatechange #ai #artificialintelligence #machinelearning #saveenvironment