Thirty years ago, at the 1992 Earth Summit in Rio de Janeiro, Brazil, the United Nations Framework Convention on Climate Change (UNFCCC) placed the importance of ‘innovative, efficient and state-of-the-art technologies’ at the centre of the response to climate change. And since then, the UNFCCC has been urging the transfer of climate technologies between countries.
Over the last decade, the pace of the development of digital tools has accelerated, and technologies designed to address climate change have become more data and AI driven.
As heads of state meet in Sharm El-Sheik, Egypt, at this year’s COP to discuss how to achieve the challenging goal of limiting global warming to 1.5 degrees above pre-industrial level, it’s important to reflect on the role that data-intensive technologies can play in the climate crisis, for what purpose and at which costs.
Considering that these tools have increasingly raised ethical concerns over the ways in which they amplify existing inequalities, we must ensure that using them to address climate change doesn’t exacerbate the inequalities that already characterise its impacts.
First, some demystification of terms. Climate action, by which we mean activities designed to reduce the impact of climate change, takes one of two forms: climate mitigation and climate adaptation.
Climate mitigation includes all activities intended to reduce greenhouse gas emissions and slow the rate of global warming, to decrease the likelihood of climate disaster. In practical terms, this might include projects that expand the use of renewable energy, improve vehicle efficiency and limit reliance on animal-based foods.
Climate adaptation, instead, includes all plans to make people and communities more resilient to climate change, such as building more effective flood defences, adapting agriculture practices to more difficult climate conditions and building new cooling systems.
So far, climate mitigation and the technologies that serve its goals have received much more attention from those involved in countering climate change. However, as it becomes increasingly clear that there currently is no credible pathway to the Paris Agreement’s 1.5 degree limit, there have been calls to take climate adaptation more seriously.
While many worry that focusing on adaptation will shift the focus away from mitigation objectives, which remain essential for warding off as many negative effects of climate change as possible, others point out that the considerable damage already caused by global warming requires us to invest in ways to make societies more resilient. For instance, Bill Gates’ climate-oriented venture capital fund, Breakthrough Energy Ventures, recently announced that it was adding a climate adaptation fund to its existing climate mitigation work.
Developing technologies of any kind for climate adaptation requires careful consideration of which communities will benefit from their use. It is widely understood that the countries with the fewest resources to address climate change are also the most vulnerable to its consequences, despite having contributed the least to its causes. But, unlike mitigation strategies, which limit emissions to the benefit of all, the effects of adaptation strategies will accrue only to those communities or countries that can afford them and implement them well, amplifying existing inequalities.
The UNFCCC makes specific demands that the countries with developed economies share their climate technologies with countries with developing economies, assuming that the former will have access to better tools and methods. However, adaptation technologies created in one context might not work and could lead to maladaptation in a different one, i.e. when a solution makes people more rather than less vulnerable to climate change. This negative outcome could be easily exacerbated when the measures implemented are data-driven.
Data-driven adaptation technologies that are developed in a particular context, with certain kinds of data infrastructures and participatory systems, may not translate well across to contexts where those mechanisms operate differently or are absent.
Instead, the countries that are most responsible for climate change should support those that are least responsible with the resources to develop adaptation technologies capable of addressing their needs and taking their capacities into account.
Two key areas of application for data-driven adaptation technologies will be city life and agriculture and, in both, we can already observe unequal distribution of the adaptation benefits.
With over 60% of the world population expected to live in cities by 2030, local authorities (as well as inhabitants) will face a number of compounding challenges. By mid-century, over 1.6 billion people in cities are projected to be exposed to extreme heat and 215 million of those will be living in a state of extreme poverty. Over 800 million people in cities are expected to be affected by rising sea levels and over 650 million to face problems of water shortage. Moreover, climate change is expected to accelerate exposure to air pollution, such as PM2.5.
In recent years, bodies such as the UN and EU have promoted data-driven solutions to these problems in the form of smart city technologies. In Seoul, South Korea, for instance, officials have identified increased air pollution and flooding as two of the top adaptation challenges that climate change will present for the city. In response to the first of these, Seoul city administration has invested in a system of air pollution sensors that measure the concentration of PM2.5 and warn citizens through mobile alerts to stay indoors or drive less when levels are high.
These kinds of air pollution sensors are increasingly used in urban environments and are a promising example of Internet of Things (IoT) technology. They do not violate personal privacy, the infrastructure and the data recorded are frequently publicly owned and open, and the system can help make the link between pollution levels and climate change explicit. But these technologies will be far more accessible in some contexts and countries than in others.
Alongside its own home intervention, the South Korean government has produced a report on the use of smart city technologies for the UN Economic Commission for Latin America and the Caribbean, encouraging Latin American countries to make use of the same tools. This is an example of the kind of technology transfer the UNFCCC recommends.
Scholars have questioned, however, whether many Latin American cities have the requisite infrastructure and underlying services that support smart city approaches in more affluent cities such as Seoul. Making use of smart city technologies in a responsible way requires access to specific hardware and a stable power grid. It also relies on government institutions with sufficient stability and resources to ensure that digital tools are implemented in ways that solve the problems they are meant to address. Passing along out-of-the-box solutions is not enough.
Another key area for climate adaptation is agriculture. A recent review found that developing countries overwhelmingly identified water and agriculture as areas where adaptation technologies are most urgently needed.
Modern farming increasingly uses AI and data-driven technologies to make food production more efficient and resilient, an approach sometimes called Agriculture 4.0 (Ag4.0) or digital agriculture.
These new systems use sensors, satellite data and historical data to make farming more precise and recommend to producers when and where to irrigate crops or use fertilisers and pesticides. As changing climate conditions make extreme weather events more common, digital agriculture can provide farmers with information about how to respond, making it an important tool for climate adaptation.
Similar technologies have also been used to create vertical gardens in urban centres. For instance, as part of the Seoul smart city tools discussed above, local authorities have encouraged companies to create so called ‘plant factories’, vertical indoor farms that rely on IoT technology to adjust relevant internal conditions. Now, the Korean company LG is planning to augment these plant factories with blockchain technology that will allow traceability and therefore improve food safety. These systems are claimed to lead to higher yields with lower inputs and pollution than conventional agricultural systems.
While digital agriculture offers some solutions to the challenges posed by climate change, it is also important to consider the inequalities it may feed into. Even relatively simple and publicly provided Ag4.0 systems, such as SMS weather alerts based on national data gathering and modelling, rely on the availability of mobile networks and phones. While this technology is widespread, it doesn’t always reach the remote rural areas where smallholders grow food.
Furthermore, the commercial systems that farmers use for operating new precision farming techniques replicate the problems of other data-driven technologies.
The data that farmers generate by monitoring crops and livestock is owned by the companies whose systems they deploy, and its use is relatively unregulated by governments. Farmers can become locked into a specific proprietary system, if the data from their farms cannot be easily migrated. Commercial suppliers of Ag4.0 algorithms are more likely to collect data and train their systems on large commodity farms, rather than smaller, more local ones, which means that their models are and will not be as relevant for these smaller farmers. The equipment and training necessary for precision farming techniques may also be out of reach for smaller farmers.
All these data-related issues can mean that large agricultural producers are more likely to be able to use digital agricultural tools in adapting to climate change. However, in regions highly affected by climate change, such as South and Southeast Asia and Sub-Saharan Africa, a significant proportion of food is produced by smallholdings of under two hectares, so the inequity of access to digital tools could mean further inequities in access to food.
As the problems generated by our rapidly warming planet compound, there is no doubt that we will need to use a wide variety of tools to adapt, and data-driven and AI tools will be part of that mix. However, as we develop these technologies, we need to proactively counter the risks of exacerbating climate inequalities. The global digital divide is still quite wide. For instance, 97% of people in developed economies have access to 4G mobile internet, in contrast to 82% in developing countries and to only 41% in the least developed countries. These differences are further amplified in rural areas.
The strategies to avoid exacerbating existing inequalities could vary: in some contexts, it might make more sense to strengthen digital infrastructure, in others it might be more useful to find low-tech or nature-based equivalent adaptations.
While technology transfer between countries has been an important feature of climate negotiations since the Rio Earth Summit in 1992, the countries which are most responsible for climate change should ensure that those least responsible have access to the resources necessary to create solutions, address their specific needs and make use of their capacities, recognising the differences in context and resources inherent to responding to the climate crisis.
The Ada Lovelace Institute, with funding from the Generation Foundation, is at the early stages of a two-year research project examining the intersection of climate change and AI. If you are interested in this work, please get in touch with the author at firstname.lastname@example.org