(AGENPARL) – gio 19 ottobre 2023 Data Science in Central Banking:
Enhancing the access to and sharing of data
Closing remarks by Alessandra Perrazzelli
Deputy Governor of Bank of Italy
3rd IFC and Bank of Italy Workshop
Rome, 19 october 2023
Good afternoon, ladies and gentlemen,
I am delighted to see so many distinguished experts gathered here today to conclude
this impressive workshop on data science in central banking jointly organized by the
Bank of Italy and the Irving Fisher Committee of the Bank for International Settlements.
Over the past three days, we have embarked on an educational journey that has
enriched our understanding of how data science is reshaping the landscape of the
financial industry and of central banking in particular. The knowledge we have shared,
the insights we have gained, and the connections we have forged are invaluable, and
they will undoubtedly play a central role in the future of our institutions and of the
financial sector as a whole.
Before I bid farewell to this event, I would like to take this opportunity to express
my sincere gratitude to all of the participants, speakers, and organizers for their
contributions.
Food for thought
Over the last three days, we have delved deep into the world of data science, exploring
its applications, challenges, and the transformative potential it holds for central
banking. We have learned about cutting-edge techniques in data analysis, machine
learning (ML), and artificial intelligence (AI) that can be applied to monetary policy, risk
management, financial stability, and many other areas that are crucial to central banks’
mandates.
Data Science is revolutionizing the banking industry. The presentations of these past
few days testify to the special and now important role of data science in crafting central
banking operations. The Bank of Italy is committed to being at the forefront of data
science innovation in the central banking community. We are constantly seeking better
ways to collect and harness new forms of data in order to improve our policy decisions
and operational efficiency. In recent years, we have started our journey through Data
Science by establishing a multidisciplinary team to consider the benefits and hidden
risks of tackling the technological challenges of artificial intelligence and machine
learning fuelled by the advances in big data, which continue to evolve at incredible
speed. In our competitive world, we are always in search of innovation. The Bank of
Italy has shown a knack for technological innovation, realizing its potential for growth
and success. In 2021, as a great way to inspire our employees and promote creativity,
collaboration, and out-of-the-box thinking, the Bank of Italy sponsored a hackathon, a
worldwide competition on applying big data, natural language processing and artificial
intelligence techniques to green and sustainable finance, that was jointly organized
with the BIS Innovation Hub of Singapore under the Italian G20 Presidency. In 2020, we
created Milano Hub, a technological hub supporting the development of innovation
and the digital transformation of the Italian financial system. After the great success of
the first two calls for proposals, we will launch a third one in the next months
In the last few years, we have redesigned our recruitment and hiring process by seeking
out new skillsets such as Big Data, Machine Learning and Artificial Intelligence. This
is just the beginning. We need to rethink our processes and foster the adoption and
development of the right skills. To achieve these goals, we have sponsored special data
science training programmes in some Italian universities. The new cohort of hires have
already started their journey, which aims to create a flatter organizational structure.
In 2020, the total amount of data created, captured, copied, and consumed globally was
around 100 zettabytes (an astounding value of 1023 bytes), and it is expected to rapidly
increase, reaching 180 zettabytes in 2025.1
The mind-boggling amount of data that is now available to us, provided we are able
to analyse it effectively, can give us a better picture of the economy at both the micro
and the macro level. The Bank of Italy has constantly striven to be at the cutting edge in
developing software and hardware platforms, enabling big data analytics2 for statistical
and economic applications.
One of the most significant takeaways from this workshop is recognizing that data is
not just a resource: it is essential for effective decision-making in central banking. The
quality, quantity, and timeliness of data can make all the difference in crafting policies
that safeguard our economies and maintain financial stability. By harnessing the power
of data science, central banks can enhance their capabilities, anticipate trends, and
respond swiftly to emerging challenges.
Data taken from https://www.statista.com/statistics/871513/worldwide-data-created/ on September
25 2023.
See, for example, ‘Big data processing: Is there a framework suitable for economists and statisticians?’,
2017 IEEE International Conference on Big Data (Big Data), 2017, pp. 2804-2811, doi: 10.1109/
Graphs.
Risks and Precautions in applying Data Science methodology
Extreme care is required when employing these new tools. Over the course of the workshop,
a number of findings have been highlighted. The ethical dilemmas that institutions and
practitioners must wrestle with in order to protect individuals from the consequences of
blindly adopting AI-based solutions are numerous and definitely profound.
Let me quickly mention some of them.
First of all, data science relies on massive amounts of information that leverage large
quantities of third-party data, and complex algorithms to generate answers. These
methodologies are often opaque and unreliable, an issue that we must address in our
work. We should always strive to pick and adopt solutions with sound and accurate
explainability.
Moreover, the kind of big data we employ typically entails selection bias because of the
peculiar characteristics of the population; increasing the number of observations does
not reduce the sampling error unless corrected for. Data should be used in a fair and
unbiased way, avoiding discrimination or harm to individuals or groups.
Second, safety and security: data and systems should be secure to protect against
unauthorized access, misuse or abuse. Human oversight is also essential in the
development and use of data and new technologies.
Third, the availability of much more detailed personal data increases the importance