Technology may solve climate change problems, but it’s also causing them

It’s that time of year again when everyone waits in anticipation for the annual fall release of the latest and greatest devices and digital applications by technology companies. Their executives are expected to pull out all the stops with hour-long showmanship-esque performances, complete with glitzy lights and spectacular marketing collateral. And the internet is abuzz with rumors about what the tech sector has in store for us. But few people are talking about the other side of the coin – the dark underbelly of the technology industry that is slowly but surely eroding our planet.

Yes, you read that right. The very industry that has given us life-changing innovations such as the smartphone and social media is also responsible for some of the most pressing environmental problems we face today. For example, a major sustainability-related tech development barely made headlines: the recent landmark EU judgment to amend the Radio Equipment Directive and mandate all mobile electronic devices to adhere to a single uniform charging port in USB-C by 2024.

The amendment is expected to reduce around 11,000 tons of e-waste in the EU alone every year. While the move is a positive step toward environmental responsibility, the unfortunate reality is that the wider technology sector is facing a sustainability crisis.

For most people, digital technology is a solution, not a driver, of our unsustainable practices. Digital applications are driving a revolution that promises and delivers transformative societal benefits. From smart cities and precision agriculture to the prospect of immersive metaverse meetings and ultra-efficient AI-enabled supply chains, the idea that digital technology will serve as a panacea for all our unsustainable practices is often touted.

Blinded by bright promises

The danger is that the promises of digital technology can blind us to its less-recognized contributions to climate woes. Emissions from the ICT sector already rival those of the airline industry. Moreover, the few studies done so far (a,b,c) suggest the ICT sector’s emissions will at least double to 2-3 GTCO₂e and if the rise in global GHG emissions is limited to 1.5˚C by 2030, contribute approximately 10% of the global GHG burden — nearly equal to that of the automotive industry. And these are cautious estimates.

This poses a dilemma not only for ICT companies, but for all technology leaders and investors across the board. It’s time for tech leaders to proactively address the sector’s sustainability challenges by critically and collectively questioning both supply and demand before its carbon footprint is set in stone, despite the near-term greenflationary impact the transition might entail. If they don’t, the secondary consequences could derail its sustainability benefits. Then, like the many industries before it, challenges could emerge in a fractured future of forced disruption.

The chains of an unsustainable digital future

In 2016, Yale environmental scientist Karen Seto and her colleagues described three types of carbon lock-ins. These lock-ins cause emissions to become entrenched in a sector and drive it along a climate-change exacerbating path. All three types of lock-ins are present in today’s ICT sector and the longer they are ignored, the more difficult it will be to reverse their consequences.

First, there are behavioral lock-ins related to customer demand for ICT goods and services. Consumers want to be perpetually connected, constantly generate and consume data, and replace their digital devices regularly — smartphone replacement cycles are now shorter than ever at under two years. Businesses, too, have locked-in behaviors, such as storing all the data they generate, irrespective of its usefulness. And businesses use complexity to track as an opportunistic reason to not even try to do so and disclose Scope 3 emissions resulting from technology use.

The overall industry, too, suffers from a subliminal, but pervasive belief that it can sidestep its own sustainability impact by delivering efficiency gains in other sectors. This, as economist William Jevons pointed out in the 1860s, is a perilous idea.

According to Jevon’s Paradox, actions that promise gains through higher efficiency are often offset, sometimes completely, by rebound effects. Jevons’ Paradoxes abound in digital technology. For instance, smart home systems, which promise energy savings of 10–25%, rarely deliver.

Second, digital technology is increasingly subject to institutional lock-ins. In this case, the sector’s governance, institutions, and decision-making affects its production and consumption, thereby shaping energy supply and demand.

Consider this: nearly 80% of all AI research is focused on the accuracy of AI, which increases its energy intensity, versus only 20% on AI’s efficiency. This bias reflects the need to correct the technology’s own deflection from sustainability and other social-good targets before employing it for the greater good.

We have found that investments in new frontiers in digital are skewed toward areas with a limited contribution to United Nations’ SDGs. To illustrate, around half of the $31 billion invested in metaverse over the past five years has been on entertainment use cases that hardly offer any sustainability benefits.

Technology legislations are largely reactionary and rarely address the need for energy sobriety. In fact, some policies lock in an entire system of antithetical climate outcomes. For example, the lack of trust in international data sharing has resulted in data protection policies that require local data storage serviced through data centers in temperate zones or powered by grids without renewable backbones.

It is tantalizing to brush off unclean power as the “electricity sector’s” problem, mitigatable by carbon offsetting; tech companies already buy more than half of all renewable offsets in the US today. Yet, each incremental renewable power taken off the grid is a net loss and takes attention away from the real target – carbon avoidance.

Third, infrastructural lock-ins, which indirectly and directly emit CO2 and shape energy supply, are evident in the ICT sector, too. The semiconductor hardware infrastructure is already locked into making devices from silicon and rare earth. Transitioning to more energy-efficient alternatives, such as gallium nitride (GaN) or biochips, will require a massive systemic change that will entail abandoning decades of investments in silicon-allied infrastructure.

Keeping the beast in digital at bay

Make no mistake, technologists can rightly claim to have delivered rapid progress in the ICT sector while managing its sustainability impact. Studies suggest the share of ICT in global emissions has stayed flat over the last decade due to breakneck innovations. But historical performance should not be expected to be indicative of future results.

We are approaching the limits of silicon hardware miniaturization, which suggests that without step-change innovation in fabrication techniques, the Moore’s Laws hardware efficiency predictions may not be realized. The low-hanging efficiency gains in network and cloud are also coming to an end as the transition from copper to optical fiber and from smaller, less efficient to hyperscale data centers near completion.

These conditions and others should serve as a clarion call for technology leaders. We cannot continue to race to optimize outcomes at unit level without a systems-level perspective. To move forward, leaders must be thoughtful about critical questions that get to the broader support of digital demand and supply, such as:

Does the demand justify the cost of fulfillment? Is there a need for technology in the first place? Is sustainability a key objective of application development? Have all the sustainability implications, including the negative effects of deployment, been considered?

How will the efficiency of supply be ensured? How will operational efficiencies across hardware manufacturing, cloud, network, and devices that enable technological applications be captured? How will the adoption of sustainability best practices and sustainability-centric design principles by all players be ensured? How can technologies, like AI-enabled sustainability solutions, be best used for the technology sector’s sake?

Lock-ins are systemic problems. Collective action involving all stakeholders will be needed to ensure the sector’s lock-ins are proactively identified and avoided. If technology leaders answer the above questions well and act now, they can ensure that digital technology does truly deliver on its many social-good promises in the future.