Despite its large carbon footprint and high amounts of waste generated every year, the tech industry might just be the solution to the world’s climate change problems. Enter artificial intelligence with all its potential, researchers and companies have leveraged on this technology to assist in finding solutions to help combat or mitigate the effects of climate change.
by Deborah Seah
When discussing climate change, many would associate the term with global warming, rightfully so. Even though both global warming and climate change are used interchangeably in conversations about harmful drastic changes on the environment, there are clear distinctions between the two.
Global warming refers to the Earth’s rising temperatures and is a result of human activities such as burning of fossil fuels, and release of carbon emissions from industrial factories. These activities cause the accumulation and trapping of greenhouse gases in the atmosphere that can cause a rise in the Earth’s surface temperature.
Climate change on the other hand does not only link to rising temperatures but also long-term changes in weather patterns, rising sea levels, and disruptions in wildlife as a result of the continued accumulation of greenhouse gases in the atmosphere.
Based on evidence by NASA, atmospheric carbon dioxide level has been increasing at an unprecedented rate over the last decades from the 1950s. This has been largely attributed to human activities and the rise of industrialisation. Global rise in temperature, warming of oceans, retreating glaciers, shrinking ice sheets, rise in global sea levels, and extreme weather events are just some of the consequences of global warming and evidence of climate change.
The proof of climate change is there, governments and private companies have also made concerted efforts to help mitigate any harmful impact current practices industries may have on the environment. Such efforts gave rise to innovative and disruptive new technologies that could change the way for example energy is produced, or materials are made to create more sustainable and environmentally-friendly options.
The technology industry is known to be one of the largest contributors to increasing carbon emissions. The Information and Communication (ICT) sector alone is estimated to have a carbon footprint of 730 million tonnes of carbon dioxide equivalents based on a research in 2018 – accounting for 1.4 percent of global emissions. This sector also uses 3.6 percent of global electricity. In a study by Belkhir and Elmeligi, the authors emphasized how the ICT sector could exceed 14 percent of the global greenhouse gas emissions by 2040. Also, the authors highlighted that smart phones alone would contribute to a major portion of carbon footprint and exceed that of other forms of hardware such as desktops, laptops, and displays.
Despite its great negative impact on the environment, the technology industry ironically could potentially help to create solutions and better understand the current climate change situation and mitigate further impact. Using artificial intelligence technologies has been aggressively studied across a wide range of fields as an innovative solution to many problems that were previously difficult or even time-consuming for humans alone to tackle.
With all its potential, it is no surprise that artificial intelligence technologies would also be included in the toolbox of researchers looking to help combat the effects of climate change. Unfortunately, tapping into this disruptive tech might not be the one to tip the scales in global carbon emissions. In a recent study by Emma Strubell and her team at the College of Information and Computer Sciences at the University of Massachusetts Amherst showed that training one NLP (Natural Language Processing) model would produce close to 300 tonnes of carbon dioxide emissions. This demonstrates the negative impact using artificial intelligence technologies could have on the environment – when ironically trying to save the environment.
Tech Companies Going Green
Big tech companies are also making their own efforts to create goals in reducing carbon emissions and combat climate change.
In early 2020, Microsoft Corp has pledged to become carbon negative by 2030 and hopes to have removed enough carbon to account for all the emissions that the company has release over the years by 2050. It has also made an announcement on 16 January 2020 that it would produce annual carbon emissions reports starting from 2020 and also sign the United Nation’s Business Ambition Pledge to limit global temperature rise to 1.5 Degrees Celsius above pre-industrial levels.
Google has also recently released its sustainability roadmap, committing to fully function on carbon-free energy by 2030. The tech company’s ambitious climate action commitments look to work with a number of partners including governments and policymakers to deploy technology to enable carbon-free energy and drive system-level change.
AWS (Amazon Web Services) has also laid out a sustainability timeline to achieve its goal of using 100 percent renewable energy usage for its global infrastructure footprint. This long-term commitment started in November 2014 and the company has since made progress in constructing renewable energy options and working together with key stakeholders and policymakers to enable more opportunities in renewable energy.
What can Artificial Intelligence do?
During a presentation at EmTech Asia 2020, Professor Tonio Buonassisi laid out a mission to his audience asking how to find the right material for a renewable energy project within a tight budget and fixed timeline. “Now how do you find a needle in a haystack?” said Professor Buonassisi.
Professor Tonio Buonassisi is a Principal Investigator at the Singapore-MIT Alliance for Research and Technology (SMART), and Professor of Mechanical Engineering at the Massachusetts Institute of Technology (MIT). He spoke at the EmTech Asia 2020 conference that was held virtually on the topic; “Addressing the Climate Crisis with Machine Learning and Automation.”
Indeed, the question is how? Professor Buonassisi then went on to highlight that over the past years acceleration of new materials development has been made possible by automation, data, and computation. The first automation has enabled high-throughput experimentation to increase outputs by even a hundred-fold. Secondly, an essential raw ingredient, data, is required for the system to function. Good data is also key when feeding machine learning systems. Thirdly, computation is the last component, this will enable decision-making and evaluation of the right materials across a large dataset.
“In every round there are thermal dynamic simulations that are fed into the machine learning algorithms. So, the algorithms have to balance the experiment which is in a sense ground truth with theory and simulation. Which can be imperfect, as there can be a gap between simulation and real-life. The machine learning algorithm has to take these two different streams of information, blend them together and make a decision about what combination of samples to try next.” Shared Professor Buonassisi, about data fusion using computation.
From this experiment that Professor Buonassisi highlighted, in using machine learning to decide on the right material, the team tested the degradation of the material after each round. The resulting degradation decreasing over the increasing number of rounds the machine learning system undergoes – demonstrating a more stable material.
Selecting the right material for developing equipment that have a positive impact on the environment is just one aspect of how artificial intelligence technologies can help in combatting climate change. Next, we look broadly as to how these innovative technologies have been researched to be used to tackle issues caused by climate change.
Artificial intelligence technologies have been used to create modelling approaches to predict weather patterns and anticipate drastic changes in the climate. A research by Catherine Buckland from the University of Oxford and her co-authors had devised a way to use artificial neural networks to determine potential dryland responses as a result of climate change and human activity. Through this method, the team was able to understand the relationship between historical data on sand deposition in semi-arid grasslands and external climate conditions, land use and wildfire occurrence. The development of such approaches using artificial intelligence technologies can help create better understanding for land use and how the landscape would change due to climate.
Surveillance strategies could also be developed to assist in analysing any risk of emerging infectious disease outbreaks that could be associated with climate change. A study led by the Public Health Agency of Canada used event-based surveillance (EBS) systems and risk modelling to help point out key risks of infectious disease outbreaks that could occur as a result of climate change. The EBS system was a combination of machine learning and natural language processing algorithms to analyse health-related internet data. Through the system the researchers were able to model the risk of disease emergence and also identify risk factors as well as any susceptible populations. From the study, such an application of artificial intelligence could be used in assisting public health decision-making to mitigate or even prevent potential infectious disease outbreaks due to climate change.
Creating an Impact
These are just a few examples of how artificial intelligence technologies can help soften the blow of climate change. Some experts say that using such disruptive technologies could be counterproductive in reducing global carbon emissions due to the amount of carbon dioxide or waste produced when developing these innovative new methods.
There would therefore be a need to tip the scale towards using more environmentally-friendly technology to reduce the tech industry’s impact on climate change, in order to achieve more a sustainable and circular economy. The use of artificial intelligence technologies to mitigate the harmful effects of environmental damage and combat climate change will bog down to the question of whether its potential positive environmental impact is worth the amount of carbon emissions that its use could produce. [APBN]
References
- NASA, Climate Change: How Do We Know? (n.d.). Retrieved from: https://climate.nasa.gov/evidence/
- NASA, Overview: Weather, Global Warming and Climate Change. (n.d.). Retrieved from: https://climate.nasa.gov/resources/global-warming-vs-climate-change/
- Malmodin, Jens & Lundén, Dag. (2018). The Energy and Carbon Footprint of the Global ICT and E&M Sectors 2010–2015. Sustainability. 10.3390/su10093027.
- Belkhir, L., & Elmeligi, A. (2018). Assessing ICT global emissions footprint: Trends to 2040 & recommendations. Journal of Cleaner Production, 177, 448–463. https://doi.org/https://doi.org/10.1016/j.jclepro.2017.12.239
- Strubell, E., et. al. (2019). Energy and Policy Considerations for Deep Learning in NLP.
- Reuters, Factbox: Big Tech and their Carbon Pledges. (January 17, 2020). Retrieved from: https://www.reuters.com/article/us-climate-change-tech-factbox-idUSKBN1ZF2E7
- Announcing Google’s third decade of climate action – our most ambitious yet. (n.d.). Retrieved from: https://sustainability.google/commitments/#
- AWS & Sustainability Timeline. (n.d.). Retrieved from: https://aws.amazon.com/about-aws/sustainability/sustainability-timeline/
- Buckland, C. E., et. al. (2019). Using artificial neural networks to predict future dryland responses to human and climate disturbances. Scientific reports, 9(1), 3855. https://doi.org/10.1038/s41598-019-40429-5
- Rees, E. E., et. al. (2019). Risk assessment strategies for early detection and prediction of infectious disease outbreaks associated with climate change. Canada communicable disease report = Releve des maladies transmissibles au Canada, 45(5), 119–126. https://doi.org/10.14745/ccdr.v45i05a02