(AGENPARL) - Roma, 28 Giugno 2024 - (AGENPARL) – ven 28 giugno 2024 Digital and green: twice the transformation, twice the win?
Closing remarks by Chiara Scotti
Deputy Governor of Banca d’Italia
Conference on ‘The macroeconomic and financial dimensions
of the green transition’
European University Institute
Fiesole, 28 June 2024
Ladies and gentlemen, good afternoon.
The conference on ‘The macroeconomic and financial dimensions of the green transition’
is now coming to a close. Let me warmly thank all the participants who have contributed
to the success of this initiative, co-organized by Banca d’Italia, the European University
Institute, the Euro Area Business Cycle Network and the Centre for Economic Policy
Research.
The value of this event lies not only in the attempt to shed light on key economic
and financial issues linked to the energy transition and the ongoing decarbonization
process, but also in gathering together academics and researchers from central banks,
other policy institutions and universities in an attempt to foster critical thinking and new
collaborations. Needless to say, the analyses of such an outstanding pool of experts are
essential in supporting decision-makers in the public and private sectors who have to
cope with the multiple challenges of the transition.
Yet while the programme has focused on many of the climate-related challenges, in my
brief concluding remarks today, I would like to highlight another challenge that I think
warrants further investigation: the twin transition, which lies at the intersection of the
green and the digital transformations.1
The digital and green twin transitions
We are living in a world where the digital transition and the green transition are
mostly taking place on parallel tracks, although some commentators have underlined
With thanks to Ivan Faiella, Luciano Lavecchia and Juri Marcucci for their valuable contributions.
some possible interactions.2 When people think about these transitions, they tend to
highlight the negative impact of the digital transition on the climate, and specifically
that the expansion in power-hungry digital technologies and services will increase their
environmental footprint.
Indeed, according to the International Energy Agency (IEA), data centres, cryptocurrencies
and Artificial Intelligence (AI) in 2022 accounted for about 2 per cent of global electricity
use. The US Energy Information Administration calculates that in 2022, cryptocurrencies
alone may have drained more than 2 per cent of the national annual power use.3
A recent study4 shows that, on average, generating an image through AI consumes as
much electricity as fully charging a smartphone, and this is on top of the energy used to
train the algorithms. This should serve as a warning, as the number of people worldwide
using these tools, even for entertainment, is increasing at an incredible rate. There is the
dreadful prospect that, in a high-demand scenario, digitalization-related energy needs
could even double in 2026.5
Clearly, these developments call for further progress in increasing the share of
low-carbon sources in the power mix,6 and for measures to improve the energy efficiency
of these digital services.7
Luckily, on the other side, there are also many good things that the digital transition can
do for the green transformation. AI tools – such as machine learning and deep learning –
can, for example, assist economists, investors, and portfolio managers in improving the
use of ESG data to build efficient portfolios8 as well as help scientists in assessing the
extent of biodiversity loss,9 map methane leaks10 or water availability.11
In what follows, I would like to mention some practical uses of digital innovation for
the green transformation that should be taken into consideration and studied further,
On the reinforcing of the green and digital transitions, see Muench, S. et al., 2022. ‘Towards a green
and digital future’.
EIA (2024). Tracking electricity consumption from U.S. cryptocurrency mining operations.
Luccioni, S., Jernite, Y., and Strubell, E. (2023). Power Hungry Processing: Watts Driving the Cost of AI
Deployment? Proceedings of the 2024 ACM Conference on Fairness, Accountability, and Transparency.
IEA 2024. Electricity 2024. Analysis and forecast to 2026.
According to the above mentioned IEA Electricity report, low-carbon electricity generation is set to
cover all global demand growth over the period 2024-26.
Zhu, H. et al., 2023. Future data center energy-conservation and emission-reduction technologies
in the context of smart and low-carbon city construction, Sustainable Cities and Society, Volume 89,
February 2023, 104322.
Lanza, A.A.G. et al., 2023. Machine Learning, ESG Indicators, and Sustainable Investment, Financial Risk
Management and Climate Change Risk, Springer Nature Switzerland.
Thompson, T., 2023. How AI can help to save endangered species, Nature, 623, 232-233.
O’Donnel, J., 2024. A new satellite will use Google’s AI to map methane leaks from space, MIT
Technology review, February 14, 2024.
Shaikh, M. and F., Birajdar, 2024. Artificial intelligence in groundwater management: Innovations,
challenges, and future prospects, International Journal of Science and Research Archive, 2024, 11(01),
502-512.
potentially also among economists like you. Although at the moment these potential
positive uses are still overshadowed by the large amount of energy required by digital
technologies, this does not mean that researchers should not push the frontier in these
areas. I will start by discussing AI-generated energy demand and supply as well as weather
forecasts12 and how they can support climate mitigation and adaptation. I will then turn to
the use of digital solutions to reinforce the green transition.
AI can help with climate transition and adaptation
AI solutions can foster the reliability of energy systems, deliver higher living standards, and
ensure safety for populations and production. The power of AI possibilities is something
we experience directly: nowadays smartphones and digital assistants provide a number
of cheap and simple home automation solutions; by coupling the Internet of Things (IoT)
with AI techniques, we can monitor and optimize the heating and cooling of our homes.
Outside of these day-to-day examples, however, there are many more ways that AI can be
seen as a ‘climate game changer’. Our smart-home solutions are just the tip of the iceberg.
Gaining a greater understanding of both when power is available and when it is needed
is crucial to ensure the efficiency and reliability of the power grid, especially once an
increasing share of intermittent renewable sources are integrated. The deployment of
smart meters to collect information, and of digital technologies such as machine learning
techniques for data analyses can lead to better forecasts of energy demand and supply,
leading to improvements in energy efficiency. In addition, by supporting ‘demand
response’, technology could minimize discrepancies between demand and supply,
reducing ‘peak loads’ and the need to add capacity to manage those peaks, and in turn
supporting the resilience and security of the energy infrastructure.
In addition to enhancing energy demand and supply forecast and management,
AI technologies can help improve weather forecasts as well as the early warning
detection of acute climate events, which are essential tools for mitigating the impact of
adverse events on populations and assets. Progress in forecasting weather patterns and
building climate scenarios could help vulnerable communities and authorities to adapt
to climate change13 and could transform climate adaptation strategies in agriculture14
(e.g. through optimized water use, early warning systems on invasions of alien species
and selection of resistant crop varieties in response to climate change scenarios) and
in the insurance sector15 (e.g. providing vast amounts of data to improve climate risk
assessment and the pricing policies for climate-related damage protection).
Bi, K. et al., 2023. Accurate medium-range global weather forecasting with 3D neural networks, Nature,
619, 533-538. Lam, R. et al., 2023. Learning skillful medium-range global weather forecasting, Science,
382, 6677, 1416-1421.
Kyungmee, K. and Boulanin, V., 2023. Artificial Intelligence for Climate Security. Possibilities and
Challenges, Stockholm International Peace Research Institute.
De Baerdemaeker, J. et al., 2023. Artificial intelligence in the agri-food sector, Scientific Foresight Unit,
European Parliamentary Research Service.
McLaughlin, M., 2022. Climate change is a major challenge for insurers, but AI and cloud can help, IBM
IBV Blog.
Digital solutions can help with climate disclosure and climate-risk identification
Other areas in which digital innovation can help with climate transition are the assessment
and selection of tools to channel funds towards green and sustainable finance, the
progress in sustainability disclosure, the collection of climate-related information, and the
evaluation and monitoring of climate-related financial risks, just to name a few important
themes. Let me briefly describe some of the efforts in these areas.
First, the BIS Innovation Hub and Banca d’Italia launched the G20 TechSprint Initiative
in 2021, a global technological competition in the form of a long hackathon designed
to show the potential for new technologies, new (big) data, and other innovations to
leverage on the potential synergies between digitalization, finance and the energy
transition.16 Another initiative in which Banca d’Italia was involved is the 2021 first call for
proposals by the Milano Hub, the Banca d’Italia innovation centre located in Milan. Some
of the projects developed under this umbrella showed how technology, and particularly
AI, can contribute to improving financial services, in terms of ESG data collection and
production.17
Second, supply chain partners, financial regulators and financial entities increasingly
require that companies, including small and medium enterprises (SMEs), provide some
kind of sustainability reporting. Digital platforms could reduce the disclosure burden.
Within this context, the United Nations Development Programme (UNDP), the Monetary
Authority of Singapore (MAS), and the Global Legal Entity Identifier Foundation (GLEIF)
proposed a framework to address SMEs’ sustainability disclosure, harnessing technology
and verified data to improve sustainability reporting while lowering costs and barriers to
reporting.18
Third, AI – particularly large language models (LLMs) – can help analysts and supervisors
in collecting climate-related information from corporates. The BIS Innovation Hub,
Banco de España, Deutsche Bundesbank, and European Central Bank launched a project
that leverages LLMs to automatically extract climate-related indicators from publicly
available corporate reports, thus enabling a widespread and cost-effective analysis of
climate-related financial risks.19
Ignazio Visco, G20 TechSprint 2021 – Presentation Event, The G20 TechSprint 2021 on Sustainable
Finance, May 2021.
One study proposed a method to address the ESG data collection problem by using adaptive AI
and natural language processing techniques, showing their effectiveness with real-world documents
(Visalli F. et al., 2023. ESG Data Collection with Adaptive AI. In Proceedings of the 25th International
Conference on Enterprise Information Systems – Volume 1: ICEIS, ISBN 978-989-758-648-4, SciTePress,
predict the energy performance certificates (EPCs) of the overall stock of Italian buildings, revealing
a discrepancy with the distribution inferred from the national cadaster (SIAPE), thereby having
significant implications for policy formulation (Braggiotti F. et al., 2024. Predicting buildings’ EPC in
Italy: a machine learning based-approach, Banca d’Italia, Questioni di Economia e Finanza, n. 850).
United Nations Development Programme, 2024. White Paper on Project Savannah: Common ESG
Metrics for Generating Digital Sustainability Credentials for MSMEs, issued jointly by UNDP, MAS and
GLEIF.
BIS, 2024. Project Gaia. Enabling climate risk analysis using generative AI.
Finally, digital platforms can support sustainable finance by integrating a vast array of
regulatory and supervisory data together with market analytics in order to support central
banks and financial supervisors to improve their monitoring of climate-related financial
risks. In 2022, the BIS Innovation Hub and the Monetary Authority of Singapore launched
a climate risk platform exactly for this purpose.20
Conclusions
The green and the digital transitions are happening whether we like it or not. We need
research to try and understand how they can reinforce each other.
One possible area of work is trying to further improve our understanding of which AI
tools and other digital developments can be used to support the green transition and
improve climate science.21 For example, AI techniques could be exploited to improve
the preparation of climate scenarios, replicating the results of conventional models, but
taking less computational time and consuming less energy.22 AI solutions could also be
used by governments to leverage massive sources of freely available climate-related
data (e.g. risk maps, satellite imagery and so on) and to generate tailored insights.23
In addition, some efforts should focus on assessing how some disruptive technologies
(such as additive manufacturing in industry, mobility-as-a-service in transportation and
AI in buildings) can help to decarbonize hard-to-abate sectors.24 In imagining future
decarbonization pathways, researchers should take into account the disruptive impact
of digital technologies on energy demand, economic growth and social development.25
Last but not least, another potential area of research is the crucial issue of how to
design incentives that could help the two transitions co-evolve and reinforce each other.
Historical evidence suggests that twin transitions tend to be lopsided: while we will
BIS, 2024. Project Viridis. A climate risk platform for financial authorities.
Rylander, Y., 2024. How will artificial intelligence influence climate science?, Stockholm Environment
Institute.
The conventional physics-based models take weeks to run and are energy-consuming: simulating a
century requires the amount of electricity used annually by the average US household. See Wong, C.,
2024. How AI is improving climate forecasts, Nature, 628, 710-712.
Peixoto, T. C. et al., 2023. Leveraging data and Artificial Intelligence for climate-smart decision-making
in government, World Bank Blogs.
For example, the Disruptive Digitalization 4 Decarbonization (2D4D) project, funded by the European
Research Council under the European Union’s Horizon 2020 research and innovation programme,
aims to carry out a comprehensive assessment of the macro-economic implications of the concurrent
digitalization and decarbonization pathways.
Creutzig, F. et al., 2022. Digitalization and the Anthropocene, Annual Review of Environment and
Resources, Volume 47, 479-509.
probably experience unceasingly fast progress in the digital world,26 the transition away
from fossil fuels will require extended periods of development.27
In this context, how can we maximize the synergies between the two transitions in
order to gradually wean future economic growth away from fossil-based energy? We
have to consider that green policies also encourage firms’ digitalization; for example,
environmental regulation as well as market-based measures (e.g. carbon pricing) promote
resource efficiency, and digital solutions are essential in achieving this objective.28
These are just a few examples, of course. What I would like you to take away is that it is
really important that the community of academics, central bankers and practitioners put
some of these topics on their research agenda so that we can do our best to maximize
the benefits in our quest for achieving twice the win in the twin transition.
Fouquet, R., and Hippe, R., 2022. Twin transitions of decarbonisation and digitalisation: A historical
perspective on energy and information in European economies, Energy Research & Social Science,
Volume 91, 102736.
“We should not ignore the experience of the past grand energy transition (from traditional biomass
energies to fossil fuels) and we should not underestimate the concatenation of challenges presented by
practical engineering, material, organizational, social, political, and environmental requirements of the
unfolding transition to a fossil carbon-free world”. Smil, V., 2024. Halfway Between Kyoto and 2050.
Zero Carbon Is a Highly Unlikely Outcome, Fraser Institute.
Fan, J. et al., 2024. Informal environmental regulation and enterprises digital transformation: A study
based on the perspective of public environmental concerns, Ecological Indicators, Volume 163.
Designed by the Printing and Publishing Division of the Bank of Italy