Climate crisis and the technology sector

The latest report (Climate Change 2023: Synthesis Report) from The Intergovernmental Panel on Climate Change (IPCC) warns that global warming has already created concerning shifts in nature, human well-being, business, and the economy in all global regions. 

The worst part? Current mitigation and reduction actions aren't enough to respond to the crisis. Immediate and deep emission reductions in all sectors are urgently needed. According to the report, limiting global warming to 1.5°C requires reducing emissions by 43% by 2030, 60% by 2035, and reaching net zero by early 2050.

Current estimates show that the Information and Communication Technology (ICT) sector contributes to approximately 4% of global CO2 emissions, comparable to the carbon emissions of the aviation sector. By 2040, projections indicate that the ICT carbon footprint could reach 14%, with data centers contributing to almost half of this growth. The entire industry must reduce carbon emissions by 45% in the next ten years to meet the goals of the Paris Climate Agreement.

It comes primarily from the consumption of electrical energy. The energy footprint of the ICT sector consumes an estimated 7% of global electricity. If it were a country, it would be the 3rd largest global consumer, behind only the USA and China.

The technology industry uses a lot of energy, particularly in running data centers, which are often powered by non-renewable sources like coal and natural gas. As innovative technologies like artificial intelligence continue to develop, the demand for energy is expected to increase even further.

The algorithms and machine learning models that power AI require a large volume of data for training and optimization. These processes consume a significant amount of energy. An interaction with ChatGPT, an artificial intelligence developed by Open AI, revealed that training large AI models can result in tens to hundreds of thousands of carbon emissions per year, equivalent to the emissions of small to medium-sized countries like Estonia. According to ChatGPT, some estimates suggest that training the AI behind GPT-3 for a year could consume the same amount of energy as several hundred households in the United States.

Technology plays a crucial, transformative role globally. However, it can become a double-edged sword if we do not incorporate sustainability early in its development.

Sustainable Code or "Green Coding" has emerged as an initiative to make software development part of the climate solution instead of part of the problem. The movement rethinks software development, making it more energy-efficient from the point it is written to testing, planning its execution in more energy-efficient environments.

Making code more sustainable is nothing more than making it more energy-efficient or less carbon-intensive.

According to the Green Software Foundation (GSF) – a non-profit organization with a mission to create a trustworthy ecosystem, standards, tools, people, and best practices for building "green" software – the three main actions that can reduce emissions in software development are:

- Using fewer physical resources.
- Using less energy.
- Using energy more intelligently.

Using energy more intelligently refers to consuming clean energy sources or electricity in a way that helps accelerate the energy transition toward a low-carbon future.

In engineering terms, sustainability translates into actions such as right-sizing resources, identifying idle workloads, and utilizing serverless architectures where possible. These efforts reduce cost and carbon in IT infrastructure, both in the cloud and on-premise.

The selection of programming languages is also essential for code performance and energy efficiency. A 2017 study from the University of Minho in Portugal showed that the fastest language is not always the most energy efficient. For example, Python used 59 times more energy than the most energy-efficient language in the ranking. The researchers concluded that no language is consistently better than others and that the situation where a language will be used is a central aspect in deciding whether it will be the most energy-efficient option.

Measuring its impact is an essential step in incorporating Green Coding into software development logic. For example, a free, open-source tool like Cloud Carbon Footprint, supplies energy and carbon emissions estimates for cloud data storage providers. GSF also provides solutions for measuring emissions, such as the Software Carbon Intensity (SCI) Specification v.1.0. This standardized protocol is used for calculating the carbon emission rate for a software system, such as calculating a carbon rate per minute for an application.

Green Coding is a crucial step toward building disruptive yet sustainable technologies in an increasingly sustainability-conscious world. In addition to contributing to reducing carbon emissions and becoming part of the climate solution, more energy-efficient software development can also result in cost reduction and increased efficiency in IT infrastructure. Additionally, adopting emissions measurement solutions can help companies identify areas where they can improve and further reduce their carbon footprint.

Being attentive to sustainability urges us to rethink our approach to software development. Small actions like choosing the correct programming language, accurately sizing resources, can significantly impact reducing carbon emissions and consolidating technology as a driver of a fairer and more sustainable world.

References
SOARES, Marcelo. The Impact of Technology on Carbon Emissions. InfoQ, 2022. Available at: https://www.infoq.com/articles/technology-carbon-impact/.
ROWE, Thomas, e VENKATESWARAN, Sunita. How Green Is Your Software?. Harvard Business Review, 2020. Available at: https://hbr.org/2020/09/how-green-is-your-software.
BLOOMBERG. How Much Energy Do AI and ChatGPT Use? No One Knows for Sure. Bloomberg, 2023. Available at: https://www.bloomberg.com/news/articles/2023-03-09/how-much-energy-do-ai-and-chatgpt-use-no-one-knows-for-sure?leadSource=uverify%20wall#xj4y7vzkg. 
WEBFORMIX. Carbon Footprint of Technology. Webformix, 2023. Available at:https://www.webformix.com/carbon-footprint-of-technology/.
GREEN SOFTWARE FOUNDATION. Green Software Foundation, 2023. Available at:https://greensoftware.foundation/. 
Energy Efficiency by Programming Language. Universidade do Minho, 2017. Available at:https://greenlab.di.uminho.pt/wp-content/uploads/2017/10/sleFinal.pdf.
GREENPEACE. Clicking Clean: Who Is Winning the Race to Build a Green Internet?. Greenpeace, 2017. Available at:https://www.greenpeace.org/international/publication/6826/clicking-clean-2017/.
OUR WORLD IN DATA. CO2 and Greenhouse Gas Emissions. Our World in Data, 2023. Available at: https://ourworldindata.org/co2-and-greenhouse-gas-emissions.

---