Walking Around the Main Gallery
Unlike its basement, the abode’s main gallery is noisy, crowded and chaotic, offering a maze-like layout we need to navigate safely. Indeed, one can easily get lost, and while checking in is pretty simple, finding one of the exit doors is undoubtedly far more elaborate. Thus, we must stand on solid ground to try to sort out the apparent mess. Based on research completed by the OECD almost 20 years ago, the UN’s International Telecommunication Union (ITU) has developed a comprehensive overview of the ICT sector framed within the context of GHG emissions. Accordingly, the sector comprises three key subsectors. 1. Manufacturing. 2. Trade. And 3. Services. While the first two are relatively easy to identify, more so if we recall our overall production blueprint, the ICT services subsector seems much more complex. Indeed, it includes software, telecommunications, data and web portals, consultancies and management, and repair services. Here, the distinction between production, distribution and exchange vanishes into thin air as the firms operating in this subsector do it all.
However, since the goal is to assess the potential footprint of the ICT sector, ITU introduces the concept of environmental life cycle assessment, following ISO standards. With this in hand, it then identifies four core categories 1. ICT end-user goods. 2 ICT network goods. 3. Data centers. And 4. ICT services. While the trade subsector has been incorporated as part of each of these categories (as is common to all of them, as is energy consumption), manufacturing has been divided in two, depending on who is the final consumer (productive consumption once again playing a significant role here). Note that IoT devices have been drafted by the end-user goods team. The services subsector has also been split in two but now excludes all activities related to entertainment (TV, Cable, broadcasting) and media (paper media, content production and all associated devices). Similarly, data centers exclude telecommunication facilities (allocated to the ICT network goods category). Finally, cable TV network operations are also assigned to the same rubric, but only if their organizational structure does not allow them to separate such operations from all other services offered. Otherwise, they are not counted.
All these nuances significantly impact how researchers have measured ICT’s environmental footprint, as we will see below. Moreover, one must wonder if the exclusion of broadcasting in the era of YouTube and Zoom is realistic.
Scoping the Gallery’s Eco-labyrinth
Capturing ICT emissions is only feasible if the companies and organizations comprising the ICT sector report them timely. ITU’s recommendations are thus a big step forward in that direction as they provide the framework facilitating such reporting from a technology perspective. However, from a more operational perspective, the Greenhouse Gas Protocol serves as the primary guidance framework. From a business perspective, corporate audit standards applied to environmental impact reporting should be part of the reporting process. Five principles permeate these proceedings 1. Relevance. 2. Completeness. 3. Consistency. 4. Transparency. And 5. Accuracy. Nothing really new under the sun here. But, of course, the issue is ensuring compliance, enforcing the principles and promoting accountability when failures occur.
Nevertheless, from a GHG emissions perspective, what is central is who owns and controls (financially and operationally) the production of ICTs and can thus report the corresponding emissions. Ownership and control can be direct or indirect, the latter occurring when emissions are produced by the activities of a business using resources owned and controlled by other firms. However, the protocol introduces three scopes, dividing indirect emissions into two categories. Scope 1 emissions are all those resulting from direct ownership and control. CO2 biomass combustion emissions and all those not covered by the Kyoto Protocol are excluded and are supposed to be reported separately. Scope 2 emissions are limited to electricity consumption by any given business -albeit actual emissions take place at the premises of the energy provider. And Scope 3 comprises all other indirect emissions stemming from extraction, production, transportation and exchange of commodities required by the activities of a firm (the so-called value chain). The GHG Protocol requires businesses to report on scope 1 and 2 emissions but makes those in the third category optional. However, that is now about to change as new reporting requirements on scope 3 emissions are on the way.
Although the various protocol scopes resemble the direct, indirect and rebound effects of ICT overall production, they are not quite the same. For starters, the GHG Protocol covers all sectors and is thus not limited to digital technologies. Second, while scope 1 emissions and ICT direct effects are similar, emissions from the other two scopes are not linked in any way or fashion to the other ICT effects. This is because the GHG Protocol scopes are essentially focused on ownership and control of the production process of all the sectors involved in producing specific goods and services – ICTs, in our case. On the other hand, ICT effects are exclusively concerned about their impact on overall production processes, regardless of who owns or controls them.
Navigating Data’s Ceaseless Tides
One could hardly go wrong by assuming that the above elaborate frameworks designed to help capture potential emissions and ecological footprints should yield reliable and robust data on ICTs. More so if we note that we can indeed combine the life cycle and scope approaches while keeping ICT effects in the mirror. Certainly, these approaches have some glaring gaps that could be eventually closed as data collection improves and reporting becomes more accurate and transparent. However, the opposite is, in fact, true. Indeed, agreement on how much ICTs emit annually has yet to be reached. We previously saw how researchers seem to pick weak indicators while ignoring, for the most part, the life cycle framework and the GHG Protocol. We also reviewed the numbers for data centers (including networks) which hover around one percent of global emissions.
The best critical overview of current ICT emissions research is here. The most comprehensive data analysis shows that in 2020 ICTs generated between 2.63 to 10.78 percent of total CO2 emissions (TV included, contra ITU’s suggestions). In absolute terms, such a percentage difference translates into almost 2,750 million CO2 tons, a number larger than the total emissions of all countries in the world except China and the U.S. That is not some marginal estimation error, for sure. However, as the review points out, access to the data is not available, so replicating calculations is not feasible. Moreover, some of the research authors work for ICT companies, raising issues of conflict of interest.
Such research builds on the life cycle approach and focuses on end-user devices, networks and data centers while thoroughly discussing the tradeoffs between efficiency and rebound effects. Unfortunately, the authors neither agree on the emissions distribution among these three sectors. We already reviewed the data centers case, which suggests that emissions in that sector are bound to increase thanks to the incoming data tsunami triggered by the often predicted IoT earthquake. In any event, ICT services are not included in any calculations. Moreover, emerging and resurging technologies such as IoT, blockchains and AI are not part of the estimation equations. And scope 3 emissions are not part of the calculations. All this suggests that ICT emissions are underestimated in all cases.
A recent report on corporate climate responsibility shows that top corporations, Big Tech included, are essentially overreporting commitments and investments in green technologies while underreporting actual emissions. The analysis uses transparency and integrity as crucial benchmark indicators, the latter defined as the quality, credibility and comprehensiveness of corporate climate responsibility. Indeed, these are pretty similar to the GHG protocol principles described above. The report hones on four critical areas. 1. Tracking and disclosing emissions. 2. Reducing own emissions. 3. Setting clear and realistic targets. And 4. Taking responsibility for unabated emissions. Most of the 25 companies assessed are well below in terms of both transparency and integrity. Of the three Big Tech companies included, Apple is ahead of the pack, while Google and Amazon are flirting with the bottom of the list. In any event, that is just another glimpse of how ICT emissions are underestimated.
But there is more. Big Tech firms (among other similarly large corporations) hold plenty of cash that sits within the financial system and thus functions as investment capital that can, of course, be directed towards fossil fuel projects. Here, such capital desperately seeks to maximize profits and will therefore move to the most rewarding ventures. A recent study has measured the impact of such investments for the first time. For example, in 2020, Apple investment capital financed projects that generated an additional 15 million CO2 tons or nearly 65 percent of Apple’s total emissions. But PayPal, a company that produces little emissions given its actual business sector and model, financed emissions 5.5 thousand times larger than its own.
Back to the Unsustainable Future
The evidence that ICTs, by default, help decrease emissions is elusive at best. And while ICTs can be part of the solution, they are also one of the main characters in the emissions game. Techno-utopia will not work here, nor will the endless and unfettered innovation hype. The lack of frameworks or methodologies to measure its environmental impact are not the problem. Ultimately, we need to take a more comprehensive approach to understand that ICTs must be produced and sustained in the long haul. Digitalization does not mean that ICT production is “immaterial.” On the contrary, their birth and growth make enormous environmental demands via sophisticated supply and value chains that remain invisible to most. The best example here is the fast-growing data centers that not only generate substantial CO2 emissions but can also have a devastating ecological impact on the geographies they inhabit.
If we really want to see ICTs as part of the solution, we must fully understand how they work vis-a-vis nature. Everything else is either noise or hype.