Why we all should consider the carbon footprint of ‘dark data’

It is estimated that we produce at least 2.5 quintillion bytes, equal to 2.5 billion gigabytes of data every day. Research suggests that more than half of all data is so-called ‘dark data’: data that is collected and stored but never used or reused. The data is stored for record-keeping, compliance purposes, possible future needs or simply forgotten about. While it may sound abstract when talking about the cloud, storing data takes up space on servers and in data centres that require electricity – lots of it.

Today, data centres consume around 1% of electricity globally, according to the IEA, but because data centre operators have prioritised energy efficiency, energy use from data centres has remained flat since 2015, despite internet traffic and data centre workloads surging. 

In addition, many data centre operators procure renewable energy and invest in carbon offsets. ICT has been the leading sector for years when it comes to corporate renewable power purchase agreements, but with more digitalisation, data consumption will reach a tipping point in the coming years, warn the Loughborough University researchers.

“It is not going to happen right now, but you will see it happening in the years to come, especially if we keep going the way we are going,” says Jackson. “Everything has sensors, governments are pushing the internet of things further, and increasingly these devices upload information into the cloud. There is a real issue with how much data we are creating each day that we don’t realise just yet.”

According to the International Data Corporation, 1.2 zettabytes of data, equal to 1.2 trillion gigabytes, was created in 2010. In 2025, that number is expected to grow to 175 zettabytes.

Many companies setting net-zero targets are just getting around to addressing all of their scope one, two and three emissions and starting to set targets for the latter. There is still little awareness of digital carbon footprints and how to tackle them, yet the Loughborough University researchers argue that reducing a company’s digital carbon footprint, or ‘digital decarbonisation’, should be a critical feature of any sustainability strategy.

The first step of digital decarbonisation is knowing and acknowledging how much data the company processes and stores, explains Tom Jackson. To help businesses get an estimate of their digital carbon footprint, Jackson and his colleague Ian Hodgkinson created a calculator based on the number of employees. For instance, a medium-sized business employing 200 people will generate 2.2 million gigabytes of data a year, of which 1.4 million could be dark data. The carbon footprint for that new data would be 4,406 tonnes of carbon dioxide annually on average.

“It sounds obvious, but a lot of organisations have no idea how much data is coming in,” says Jackson. “So, one thing a company can do is start looking at the process of how data is coming in and is reproduced, who is using that data, how much is stored and what the value of that data is to the business.”

One way of reducing a company's digital footprint is rethinking the way we use knowledge. Instead of having employees look up the same thing multiple times, this theory builds on the idea that sharing and reusing knowledge between employees could lead to a reduced carbon footprint.

“We are moving on to a society where we have ‘plastic knowledge’," Jackson explains. "We go to a search engine, we roughly figure out what we wanted to know or how to do something, and then we discard that information. So, it is kind of like a plastic bag that you use once and then throw away. When other people then keep looking up the same information, that creates dark data, and potentially it stifles innovation within the company.”

Single-use knowledge means there are high risks of this knowledge being forgotten or lost over time, for instance because of staff turnover, system upgrades and poor knowledge capture. If knowledge is reused and shared with colleagues, it will be combined with their own and not only avoid 'dark data' but also potentially create new insights and ideas, argue the Loughborough researchers.

On an individual level, there is only so much we can do. Deleting a few old photos from the cloud will not make a big difference when we are surrounded by technical systems that are all generating data. However, like with companies, just being aware of how big your digital footprint is could be the first step. Most tech companies let you download your personal data. For instance, Google Takeout allows you to download a copy of everything Google knows about you. While transferring the data will create additional carbon, it will give you an idea of what kind of information is being generated and stored about you. 

“While it is difficult to change anything as an individual, we do need to rethink the amount of data we are generating as a society,” says Tom Jackson. “We need to think about it the same way as we have been thinking about (recycling) plastic; do we really need to have all these apps and devices that are generating and storing all this data?”

Solutions targeting dark data and digital decarbonisation are not yet straightforward. For instance, if we start to automatically remove old or unused data, how do we decide what information is important to keep, and what is not? And would making it more expensive to buy storage exacerbate a digital divide between poor and rich? If giving everyone ownership over their personal data could be a way to reduce emissions, how do we keep that data secure and who oversees it?

For Jackson, it is important to start thinking about necessary changes now because the issue will only get bigger over time.

“I don’t want to get down the line a decade from now where we say: ‘Oh, we didn’t see this coming’. We need governments to start thinking about it now and plan for the future. Data offers a way forward, but to make that happen sustainably, we need to change the way businesses work.” 

This article originally appeared in our sibling publication Energy Monitor.

// Main image: Laptop left in room. Credit: Aleksey Boyko via Shutterstock