(PlUA/Shutterstock)
Earth Day is today, and millions of people are taking the time to reflect on how their actions impact the planet. These reflections are especially important for IT professionals, as every byte of data processed increases our collective carbon footprint. Data of the unstructured kind seems to have a greater environmental burden, especially when it is used to power AI initiatives.
According to the US Energy Information Administration (EIA), data centers in the United States consume 1.8% of all electricity produced in the country. A May 2021 paperPublished in Environmental Research Letters. Whats more, these American data centers–which account for 20% of all the data centers in the world–also consume a large amount of fresh water, they found.
However, not all data centres are created equal. AWSThe company has pledged to use 100% renewable energy by 2025, and to reaching net-zero carbon emissions by 2040. It claims that its large public data centers have a smaller footprint than private data centers owned by individual companies.
The company claims that companies who move to AWS Cloud from their existing infrastructure can reduce their carbon emissions by 88%. This is because our data centers offer environmental economies of size. It claims that organizations use 77% less servers, 84% more power, and get 28% more solar and/or wind power from the AWS Cloud than their on-premises data centers.
The Google Cloud data center near Delfzijl in the Netherlands (Rudmer Zwerver/Shutterstock)
AWS launched a brand new service last month. Carbon footprint toolCustomers can now see the impact of IT operations on emissions. This brings it up to date with its competitor. Google CloudThe tool was similar to that launched by Google Cloud last October. Google Cloud has committed to using 100% renewable electricity in its data centers by 2030, according to the company.
Different types of data have different environmental impacts just as different data centers can have different impacts. Unstructured data like video, audio, and images takes up more storage and processing time than structured data like the tabular data you would find in a relational database. Semi-structured information, such as JSON data, that you would find within a NoSQL Database, falls somewhere between.
Data automation and intelligence firm AparaviIt is asking companies to reevaluate not only how many data they store and process but also how much unstructured and unstructured data they have. Unstructured data releases more carbon into our atmosphere.
Adrian Knapp, founder and CEO of Aparavi, said that the greater the company’s unstructured data, the greater the data footprint. Companies can go green by identifying redundant and obsolete data (also known as ROT data). Unaccounted data can be detrimental to the environment because it takes up space on servers, slows down processing and is inefficient.
Similar themes are emerging from Datadobi, a provider for unstructured data management software. Michael Jack, the co-founder and chief operating officer of Datadobi, believes that unstructured information is key to achieving environmental, sociojustice, and governance (ESG), initiatives.
Jack says that in honor of Earth Day, I want enterprises to remember that unstructured data is an important part of the ESG conversation. ESG can be approached holistically by encouraging and enabling enterprises to move away legacy models, which are storing data in a digital landfill that is taking up space, money, precious resources, and providing very little return.
Today’s massive AI initiatives often use unstructured data. Companies can automate actions using imagery or text from a neural network architecture. This form of AI is becoming more popular, and so does its environmental impact.
AI usage is growing year-on-year in many sectors. This is especially true for financial services and insurance, where the rate at which AI adoption is increasing is more than 37% per year, according to KPMG.
Linh C. Ho is the CMO of at Zelros, which creates AI solutions for insurance firms. This has a significant environmental impact and will not be decreasing anytime soon.
Due to the COVID-19 Pandemic that closed schools and factories and eliminated flights, the global carbon footprint dropped by 6% in 2020. The world’s carbon footprint is increasing as the global economy is recovering.
There is one bright spot in all this: the potential to use advanced technologies not only to monitor carbon emissions at a finer level but also to manage them better. ForresterAnalyst Abhijit Sulil is optimistic that this will have a significant impact.
Sunil believes that the future will see more use of emerging technologies such as AI/ML, blockchain, IoT, and automation in sustainability applications. They can be used to reduce carbon emissions, improve measurement and accounting, and reduce waste.
Sunil warns that relying on emerging technologies can pose a risk because they are computationally expensive and energy-intensive.
He also warns that they could increase carbon emissions at the edges and increase e-waste. As solutions and services providers look to integrate these technologies with sustainability, it is important that we balance the risks associated with them as they mature, especially when there are less carbon intensive alternatives. It is crucial to identify the right use case, scale, and scope in order to make technologies more beneficial than dangerous for climate action.
Similar Items
Mobility Data reveals Climate Change Pathways for a Post COVID-19 World
European Cities Use AI to Reduce Carbon Footprints
How machine learning can help us stop climate change
