Presenters

Source

Is “IT for Green” Truly Green? Unpacking Tech’s Ecological Impact with Gauthier Roussilhe 💡🌍

How many times have you heard the bold claim: “Every time we emit one ton of CO2 in tech, we save 10 tons elsewhere?” This statement, often repeated across the tech industry, sounds fantastic. But is it true? And how do we really assess the net positive impact of digital technologies on our environment?

Welcome to a deep dive into “IT for Green,” where we challenge assumptions and explore the complex reality behind tech’s environmental footprint. We recently sat down with Gauthier Roussilhe, Head of Research and Development at Ublau and a former PhD researcher at Melbourne’s RMIT University, to unravel these questions. Gauthier, a strong voice against the environmental footprint of digital technologies, shared his invaluable insights from an in-depth study commissioned by ADEME, the French agency for ecological transition.

The Myth of Effortless Green: Why We Need a Reality Check 🤔📊

Gauthier’s journey into “IT for Green” began during the COVID-19 pandemic, sparked by that very statement: “one ton of CO2 emitted by the tech sector is 10 tons avoided in other sectors.” He recalled a meeting where telecom operators and IT companies used this to justify the sector’s footprint, claiming it positively contributes to France’s transition. Curious, Gauthier spent two months in 2021 diving into major industry reports from organizations like GSMA and GI. His conclusion? Most numbers were “very, very, very flawed” and unsuitable for decision-making.

This disconnect between industry narratives and scientific rigor prompted ADEME to launch its groundbreaking study, IT for Green: Environmental Assessment of the Direct and Indirect Impact of Digital Technologies for Some Use Cases. This was a world-first: a state-commissioned report independently assessing tech’s impact on its own territory, rather than relying on industry-driven analyses. This shift in perspective, from a company looking at its solution to a state evaluating tech’s contribution to national carbon reduction targets, profoundly changes the results.

Diving Deep: The Remote Work Paradox 🏠💻

Among the nine use cases studied, Gauthier found assessing remote working in France particularly fascinating. This topic, linking technological development with ecological impact reduction, dates back to 1976 with the father of remote working, Neils, who claimed it would reduce CO2 emissions.

The Consequential Analysis Approach 🗺️⏳

To understand remote work’s true impact, Gauthier and his team adopted a consequential analysis methodology. Unlike traditional life cycle assessments (LCA) that allocate impacts, consequential analysis asks: “What will be the effects, both expected and unexpected, of deploying this technology in a specific context?”

Their study spanned from 2022 to 2035, focusing on three types of French urban areas (small, medium, large, excluding Paris). They designed three scenarios:

  1. Business As Usual: Remote working continues as it is today.
  2. Massification: Increased remote working days and intensity, especially for categories not currently teleworking much.
  3. Reduction: Drastically decreased remote working, reflecting employer preferences.

Crucially, they considered the “reference scenario” – what would happen otherwise. Since remote working already exists, they assessed the impact of changes in its prevalence.

Mapping the Ripple Effects: Complexities Unveiled 🕸️🤯

The team meticulously mapped out the consequences of remote working:

  • IT Solutions: Surprisingly, the environmental footprint of collaboration tools like Teams and Slack was largely neutralized in the analysis. Why? Because these tools are already widely deployed, whether people work remotely or not. Their presence in all scenarios meant no significant differential impact.
  • Car Travel Substitution: This is the biggest expected effect. Remote work reduces commuting, avoiding CO2 emissions from internal combustion engines. However, this substitution is not complete. Daily travel patterns shift from linear (home-office-supermarket) to “star-shaped” (home-pharmacy-school-supermarket), often using less carbon-intensive modes like bikes. Also, remote workers might compensate by taking longer leisure trips or weekend getaways, partially offsetting daily savings.
  • Home Heating & Cooling: If you work from home, you might heat your home more. The study assumed homes are heated to 19°C when occupied and 16°C when empty. However, whether this adds impact depends on individual habits – if you don’t lower your heating when away, remote working has no additional heating impact. Scientifically, six extra hours of home occupation only account for two more hours of heating on average.
  • Office Building Energy: Counterintuitively, less human presence in offices often had no significant impact on heating, cooling, or ventilation in open-plan spaces. A single employee requires the same heating as ten. In fact, fewer humans mean less ambient heat. The long-term impact focused more on the degrowth rate of office construction.

The Underwhelming Truth: Marginal Gains 📉🤏

After running the complex model, the findings were clear: massification of remote working in France was net positive on CO2 emissions across all urban areas. However, the impact was marginal.

  • The study found roughly 1,000 tons of CO2 avoided per year per city.
  • Over 14 years, even a large city like Marseille avoided only about 40,000 tons of CO2 emissions.
  • This contributed only 2% of the annual effort needed to reduce car travel emissions within cities.

Furthermore, these positive and negative impacts are decreasing over time. As cars become more electric and homes more energy-efficient (e.g., heat pump deployment), the potential for remote work to avoid emissions diminishes. This “diminishing returns” is actually a good sign, indicating progress in the broader energy transition.

The study also looked at 16 indicators beyond CO2, excluding biodiversity. While most showed similar dynamics, the depletion of minerals and metals consistently showed a negative impact, as these resources are non-renewable.

A crucial insight: if France were to reduce its current level of remote working, it would actually add emissions. So, while the impact is marginal, having remote work is better than not having it in the current context.

The “Golden Case” That Wasn’t Quite Gold: Dynamic Line Rating (DLR) ⚡️🌬️

Gauthier then shared another use case, which he called a “golden case” for positive impact: Dynamic Line Rating (DLR) for electricity transmission lines in France. This solution, managed by the state-owned monopoly RTE, seemed perfect.

  • Technology: Connected weather stations allow dynamic adjustment of transmission line capacity, rather than static, seasonal values. This prevents energy loss and increases the amount of wind-generated electricity that can be injected into the grid.
  • Impact: DLR adds 300 megawatts per hour of low-carbon wind power per line annually. This is great, but 30% of this new electricity stays in France, while 70% is exported, mainly to Belgium, Netherlands, and Germany. It also avoids the impact of civil works for line upgrades.
  • Rebound Effect: The study even considered a known rebound effect where people using green energy sources tend to consume more.

The Stark Reality: Still Marginal 📉🤏

The solution showed a net positive impact on most indicators (excluding minerals/metals). Over the study period, it avoided 33,000 tons of GHG emissions.

However, the sheer scale of the challenge once again dwarfed the achievement. By 2035, only 50 lines in France are projected to be equipped with DLR. Compared to the annual effort needed to decarbonize France’s electricity production, DLR contributed a mere 0.08%.

While the footprint of the IT equipment (e.g., 60 kg weather stations) was negligible compared to the solution’s overall effect, Gauthier expressed disappointment. The problem isn’t that the solution isn’t good; it’s that even in a “perfect setup” with a monopoly, stable funding, and no human behavioral dependencies, the impact remains incredibly small relative to the enormous decarbonization targets.

The Elephant in the Room: Enabled Emissions & Technologist’s Dilemma 🐘💼

Gauthier highlighted critical challenges often overlooked in “IT for Green” discussions:

  • Counteracting Solutions: Digital solutions can negate each other. For example, promoting ride-sharing (increasing car occupancy) can reduce the avoided emissions from remote working. We cannot simply assume that aggregating digital solutions automatically leads to cumulative positive effects.
  • The Oil & Gas Connection: The biggest unaddressed issue might be enabled emissions from the digitization of the oil and gas sector. Digitalization isn’t just for “nice and friendly” sectors; oil and gas companies are massive purchasers of IT solutions. Gauthier’s microeconomic analysis suggests four times more money flows from ICT to oil and gas than to renewable and nuclear energy. This means tech might be inadvertently contributing to increased carbon intensity and prolonged reliance on fossil fuels, as famously highlighted by Will and Holly Alpine’s experience at Microsoft.

This raises a profound question for technologists: Where do we draw the line of responsibility?

Rebound Effects: Not Always a Villain, But Always a Consideration ♻️💡

Gauthier clarified that rebound effects are not automatic. They depend heavily on context. For example, a city with a decreasing budget might not reinvest savings from efficiency gains, thus avoiding a rebound effect.

He used the Vinted case as an example: the platform enables a secondhand market, but 30-40% of the money users earn is often reinvested in new clothes, frequently from ultra-fast fashion brands like Shein. This is a negative rebound effect. Gauthier suggested a solution: if Vinted paid users in its own digital currency usable only on the platform, it could “close the loop” and prevent external negative reinvestment, though this might make the economic model less attractive.

However, rebound effects can also be positive. An e-bike study in Oslo found that people primarily use e-bikes to substitute car travel. Furthermore, e-bike users tend to travel three times the distance they would on a regular bike – a 300% rebound effect. But since this growth is in a decarbonizing mode of transport, it actually reduces emissions.

Context is King, Not the Solution Itself 👑🌍

Gauthier’s most powerful takeaway: “Context is key, not a solution.”

Remote working itself has no inherent impact; it’s remote working in a rural area versus a massive city that produces impact. The solution’s effectiveness is tied to the specific infrastructure, regulations, and societal habits of its deployment. The findings for France, for instance, might be vastly different in a country like China due to varying policy contexts and energy mixes.

Technologists must consider:

  • Where do I want to apply my efficiency gains?
  • How do I control for rebound effects, both positive and negative?
  • Am I making an “unsustainable” industry more efficient, thereby prolonging its existence?

Not every company or solution needs to have a net positive impact. Climate mapping or satellite solutions, for example, might add a small amount of emissions but provide invaluable data for other sectors to decarbonize.

A Glimmer of Hope: Looking East for Inspiration ✨🌏

Despite the complexities, Gauthier ended on a positive note. He highlighted the “massive carbon emission reduction” recently seen in countries like China and India. China, for instance, reduced its carbon footprint by 300 megatons in one semester. While these nations are currently among the largest emitters, they are also decarbonizing the fastest. Pakistan’s rapid adoption of solar panels, driven by the unreliability of its traditional grid, is another example of unexpected, swift progress.

This journey with Gauthier Roussilhe underscores a critical lesson: the path to a greener digital world is far from straightforward. It demands rigorous assessment, an understanding of complex ripple effects, and a keen awareness of context. For technologists, it’s not just about building solutions, but understanding their true, multifaceted impact in the real world.

Appendix