SUNANDO ROY – On Banking, Finance and Society

As central banks navigate an increasingly complex financial landscape, Artificial Intelligence (AI) is emerging as a transformative tool to enhance economic stability and policymaking. The recent BIS report ( January 29, 2025) on AI governance in central banks provides a comprehensive framework for integrating AI into financial institutions while mitigating its associated risks. ( https://www.bis.org/publ/othp90.pdf ). This report provides specific suggestions on how central banks can identify, analyse, report and manage risk associated with the adoption of AI models and tools in their organisations. The report argues that central banks need to find a balance between fostering innovation using AI and mitigating the different risks that this technology may generate. Good governance schemes for the adoption of AI in the organisation, with a holistic view beyond technology, might help to achieve such a balance. The BIS report clearly suggests a need based , jurisdiction-specific timeline and strategy for AI adoption. This is what BIS terms “ an adaptive AI governance framework.”

The well thought-through recommendations of the above report remind me of the story from Greek mythology, the bed of Procrustes. The story goes like this – Procrustes was a villainous figure who lived along the road from Athens to Eleusis. He was a blacksmith or an innkeeper who offered travelers a place to sleep in his special iron bed. However, Procrustes had a twisted sense of hospitality—he made sure that every guest fit the bed exactly, but in the most gruesome way. If the traveler was too tall, he chopped off their legs to make them fit.If the traveler was too short, he stretched them on a rack until they matched the bed’s length. This horrific practice made Procrustes infamous, and his name became synonymous with forced uniformity and cruelty. Eventually, Procrustes met his match in Theseus, but that’s another story.

The takeaway of the story, and in the above BIS report is that the procrustean bed of AI adoption may not fit all central banks and an adaptive framework may be needed based on jurisdictional specificities. This is particularly true for emerging market central banks. For emerging market central banks, where economic volatility, regulatory gaps, and resource constraints pose significant challenges,AI presents both unique opportunities and vulnerabilities. The key question is: how can central banks in emerging and frontier markets leverage AI to strengthen economic stability while ensuring responsible and ethical adoption?

How to fit AI in Central Bank mandate

AI’s application in central banking is multifaceted, spanning areas such as economic forecasting, risk management, financial supervision, and payment systems. The ability of AI to analyze vast datasets and detect patterns can enhance monetary policy effectiveness and improve real-time decision-making. The BIS report highlights that “Central banks adopt AI to enhance efficiency, improve operational robustness and inform decision-making in different areas of the organisation” (Bank for International Settlements, 2025, 4). For emerging market central banks, where access to high-quality economic data may be limited, AI-driven nowcasting can improve GDP and inflation forecasts, reducing the lag between policy formulation and implementation.

In financial supervision, AI can automate the detection of systemic risks and financial irregularities, providing regulators with deeper insight into market behavior. The report underscores AI’s potential in regulatory compliance, stating, “AI tools help assess and monitor financial institutions, covering aspects like credit risk, exposure and portfolio risks” (Bank for International Settlements, 2025, 6). This is particularly valuable for emerging markets, where financial institutions may be more susceptible to credit cycles, liquidity risks, and compliance challenges. By adopting AI-driven risk assessment models, central banks can improve financial oversight and ensure greater stability in banking systems.

Artificial Intelligence Adoption in Central Banks: An Emerging Market Perspective

Let’s explore what’s happening beyond the domain of advanced country regulators. There too, Artificial Intelligence (AI) is revolutionizing central banking by enhancing efficiency, improving macroeconomic forecasting, detecting financial anomalies, and strengthening regulatory oversight. With the rapid digitalization of financial markets and the increasing availability of structured and unstructured data, central banks worldwide are leveraging ( in widely varying degrees) AI to optimize operations. However, AI adoption also brings concerns related to transparency, data privacy, algorithmic bias, and financial stability risks. While central banks recognize AI’s potential, navigating complexities of integrating these technologies in a responsible and sustainable manner has proved challenging and often acted as a deterrent ( playing it safe by sticking to status quo).

One of the most significant applications of AI in central banking is macroeconomic forecasting. AI models can analyze vast amounts of economic data to provide more accurate and timely projections. For instance, the Bank of Indonesia utilizes machine learning algorithms to assess the impact of foreign investor activities on exchange rates and monetary policy. Another key application is anomaly detection in financial transactions and reporting. The Bank of Canada has adopted AI-driven solutions to identify irregularities in data submitted by financial institutions, helping to improve regulatory compliance and risk management.

Central banks are also using AI to analyze unstructured data sources such as news articles, social media trends, and corporate reports. The Central Bank of Malaysia, for example, has implemented AI-powered natural language processing (NLP) models to analyze hundreds of thousands of news articles, allowing for better economic forecasting and demand-side analytics. AI is further playing a crucial role in regulatory compliance and fraud detection, with machine learning models enabling real-time identification of suspicious transactions and financial misconduct.

The benefits of AI adoption in central banks are extensive. AI enhances efficiency by automating routine tasks, allowing financial authorities to focus on strategic decision-making. It strengthens economic forecasting, providing deeper insights into financial trends and macroeconomic indicators. Moreover, AI-driven risk management tools help central banks detect financial vulnerabilities and intervene proactively to prevent potential crises.

The Reserve Bank of India (RBI) has recognized the transformative potential of AI while also acknowledging the associated risks and regulatory challenges. The RBI has adopted a cautious yet proactive approach to AI integration, focusing on responsible and ethical AI use in the financial sector. In December 2024, the RBI launched the Framework for Responsible and Ethical Enablement of Artificial Intelligence (FREE-AI) initiative. A high-level committee has been tasked with assessing the level of AI adoption in India’s financial sector, reviewing global best practices, identifying AI-related risks, and proposing governance guidelines for responsible AI implementation. The committee’s primary objective is to develop a regulatory framework that ensures AI-driven financial services remain secure, unbiased, and aligned with the RBI’s overarching goals of financial stability and consumer protection. The recent BIS framework provides a best practice guidance for the team.

The RBI is actively exploring AI applications in banking supervision and fraud detection. AI-powered tools are being tested to enhance regulatory compliance monitoring, identify fraudulent transactions, and improve the detection of unusual financial activities. By leveraging AI for real-time analysis of financial data, the RBI aims to strengthen its oversight capabilities and improve its response to emerging risks in the banking sector. Additionally, AI-driven models are being utilized to track macroeconomic indicators more efficiently, helping the RBI in formulating monetary policy decisions based on advanced data analytics.

Looking ahead, the RBI is expected to expand its AI strategy to further enhance risk management, cybersecurity, and fintech regulation. One of the key priorities will be implementing AI-driven models for systemic risk assessment, enabling early detection of financial crises. Additionally, AI-powered cybersecurity tools are likely to be integrated into the RBI’s framework to enhance fraud detection and financial transaction security.

Despite its advantages, AI integration presents several challenges in the emerging market financial regulation and supervision. One of the primary concerns is data privacy and security, as AI models require access to large datasets that may contain sensitive financial information. Algorithmic bias is another significant risk, as AI systems may inherit biases from the data they are trained on, potentially leading to unfair financial assessments or discriminatory policy recommendations. Overreliance on AI also introduces operational risks, as system failures, cyber threats, or incorrect model predictions could disrupt financial stability. Additionally, the lack of standardized regulatory frameworks for AI in banking creates uncertainty regarding accountability and oversight.

AI’s role in digital currencies is another area of growing interest. The SUERF Forum states that AI is becoming increasingly influential in shaping digital payment strategies, from enhancing security to assessing economic impacts. Meanwhile, discussions at the Central Bank AI Conference emphasize the need for global cooperation, with experts noting that AI in central banking is still in its early stages and that ongoing dialogue is essential to harness its full potential responsibly. The BIS Innovation Hub is one such initiative, bringing

Overall, adoption of AI in central banking is growing at an accelerated pace.As AI technologies continue to evolve, central banks must strike a balance between fostering innovation and ensuring responsible governance. Establishing robust regulatory guidelines, improving AI transparency, and mitigating systemic risks will be crucial to ensure that AI enhances central banking operations without compromising financial stability. The recently released BIS report will be an important benchmarking tool in this respect.

Voices of Concerns

The Bank for International Settlements (BIS) highlights AI’s transformative impact on the economy, cautioning that its benefits must not come at the cost of systemic risks: “Artificial intelligence is transforming the economy, but central banks must ensure its benefits do not come at the cost of systemic risks.” The European Central Bank (ECB) similarly recognizes AI’s role in improving forecasting models, stating that AI-driven models provide deeper insights into macroeconomic trends but must be used responsibly to avoid over-reliance on automation. The Bank of Canada further warns against biases in AI models, emphasizing the importance of transparency in AI-based decision-making for inflationary trends and financial risk assessments.

Ethical and regulatory concerns are also central to AI implementation in central banking. The Bank of England stresses that AI use must be built on public trust, advocating for transparency and fairness: “A central bank’s use of AI must be trusted—transparent, fair, and aligned with public interest.” The Frankfurt School of Finance & Management warns that while AI could optimize monetary policy operations, safeguards must be in place to prevent unintended economic consequences. Similarly, researchers from the Journal of Banking and Finance discuss how AI’s speed and complexity could amplify financial instability if left unchecked.

Thus, while AI has the potential to transform central banking, institutions must ensure that its implementation prioritizes stability, transparency, and public trust.

Challenges in AI Adoption

Despite its promise, AI adoption in central banks is fraught with governance challenges, cybersecurity threats, and data integrity risks. One of the primary concerns highlighted in the BIS report is the risk of data security breaches and AI-driven misinformation. The report warns that “GAI outputs may contain confidential data, leading to security breaches, unauthorised access or privacy violations” (Bank for International Settlements, 2025, 15). In emerging markets, where cybersecurity frameworks may still be developing, central banks must be particularly cautious in deploying AI models that process sensitive financial data.

Another challenge is the dependence on third-party AI providers, which can limit central banks’ control over critical financial models. The report raises concerns about third-party risks, stating, “Dependency on third-party providers means that central banks cannot modify or adjust the technologies according to the evolution of their business needs” (Bank for International Settlements, 2025, 9). For emerging markets, reliance on external AI solutions may expose central banks to vendor lock-in, data sovereignty issues, and compliance risks. Ensuring that AI models align with domestic financial regulations and national security interests is crucial for long-term stability.

Furthermore, capacity constraints pose a significant hurdle in AI adoption. Many central banks in emerging economies face limited technical expertise and resource constraints, making it difficult to develop in-house AI capabilities. The report emphasizes the importance of workforce training, noting that “The proper functioning of central banks largely depends on the specialisation and expertise of their staff” (Bank for International Settlements, 2025, 23). Without a skilled workforce to oversee AI implementation, central banks risk misinterpreting AI-driven insights or relying on flawed models that could undermine financial stability.

Strategies for Responsible AI Implementation

To maximize AI’s benefits while mitigating risks, central banks must adopt a structured and adaptive AI governance framework. The BIS report recommends a cautious approach, advising that “A safe AI adoption could cover the following domains: governance; legal and compliance; information security and privacy; cyber security; third-party risk management; business continuity; and other operational risks associated with the level of digitalisation and exposure of the organisation” (Bank for International Settlements, 2025, 24).

First, central banks must develop AI governance structures that align with national economic priorities. This includes establishing AI oversight committees to regulate adoption, monitor risks, and ensure ethical AI use. Second, enhancing cybersecurity resilience is essential. AI-driven financial models should be integrated with strong encryption protocols, real-time monitoring systems, and multi-layered security defenses to prevent cyber threats and data breaches.

Third, central banks should prioritize domestic AI capacity building through training programs, research collaborations, and partnerships with international financial institutions. This will reduce dependence on foreign AI solutions and enable central banks to develop context-specific AI models tailored to their economies. Lastly, a phased AI adoption strategy—where AI is first implemented in non-critical banking functions before expanding to core financial stability operations—can help mitigate unintended risks.

Conclusion: A Balanced Approach to AI in Central Banking

AI offers emerging market central banks a powerful tool for improving economic stability, but its adoption must be strategic, well-governed, and risk-aware. The BIS report provides a roadmap for central banks to harness AI’s benefits while ensuring robust governance structures are in place. As the report highlights, “Good AI governance is important not only for complying with national strategies, laws or regulations but also for ensuring the alignment of AI uses with the organization’s strategy and objectives” (Bank for International Settlements, 2025, 21).

By adopting AI with a clear governance framework, investing in cybersecurity resilience, strengthening regulatory oversight, and building AI expertise, emerging market central banks can leverage AI as a force for economic stability and financial resilience. The key to success lies in balancing innovation with responsibility, ensuring that AI-driven financial decision-making enhances stability rather than introducing new risks. Through a well-managed AI strategy, central banks can future-proof their financial systems and contribute to sustainable economic growth in an era of rapid digital transformation.

Reference

Bank for International Settlements. Governance of AI Adoption in Central Banks. BIS Representative Office for the Americas, January 2025. https://www.bis.org/publ/othp90.pdf

Bank for International Settlements. Artificial Intelligence and the Economy: Implications for Central Banks. BIS Annual Economic Report, June 25, 2024. Accessed January 29, 2025. https://www.bis.org/publ/arpdf/ar2024e3.htm.

European Central Bank. Artificial Intelligence: A Central Bank’s View. Speech by Christine Lagarde, July 4, 2024.

https://www.ecb.europa.eu/press/key/date/2024/html/ecb.sp240704_1~e348c05894.en.html.

Macklem, Tiff. Artificial Intelligence, the Economy and Central Banking. Bank of Canada, September 20, 2024. Accessed January 29, 2025. https://www.bankofcanada.ca/2024/09/artificial-intelligence-the-economy-and-central-banking/.

Benford, James. TRUSTED AI: Ethical, Safe, and Effective Application of Artificial Intelligence. Bank of England, September 2024. Accessed January 29, 2025. https://www.bankofengland.co.uk/speech/2024/september/james-benford-speech-at-the-central-bank-ai-inaugural-conference.

SUERF – The European Money and Finance Forum. AI in Central Banking. Accessed January 29, 2025. https://www.suerf.org/events/ai-in-central-banking/. ( Downloadable presentations available)

Jacobs, Julian. “How Central Banks Are Already Deploying Artificial Intelligence.” OMFIF, September 6, 2023. https://www.omfif.org/2023/09/how-central-banks-are-already-deploying-artificial-intelligence/.

Shukla, Saloni. “RBI Announces ‘FREE-AI’ Committee to Develop AI Framework.” The Economic Times, December 26, 2024. https://economictimes.indiatimes.com/news/economy/policy/rbi-announces-free-ai-committee-to-develop-ai-framework/articleshow/116684195.cms.

Appendix : Artificial Intelligence in Central Banks – A Chronology

Appendix: Chronological Summary of Works on AI in Central Banking (2010-2025)

2010

• June 2010: The Bank of England published a discussion paper titled The Use of Artificial Intelligence Techniques in the Financial Sector, exploring early applications of AI in financial services and potential implications for central banking.

2012

• March 2012: The European Central Bank released a working paper, Machine Learning for Credit Risk Assessment, analyzing the potential of AI models to improve credit risk evaluation processes within central banks.

2015

• November 2015: The Federal Reserve Bank of New York hosted a conference on Big Data and Machine Learning in Macroeconomic Analysis, discussing the integration of AI techniques in economic forecasting and policy analysis.

2017

• April 2017: The Bank for International Settlements (BIS) published a report titled Big Data in Central Banks, examining the role of AI and big data analytics in enhancing central banking functions.

2018

• September 2018: The International Monetary Fund (IMF) released a staff discussion note, The Economics of Artificial Intelligence, exploring the macroeconomic implications of AI adoption, including considerations for monetary policy.

2019

• December 2019: The Bank of Canada published a paper, Artificial Intelligence in Banking: The Changing Landscape, analyzing the adoption of AI in banking and its potential impact on financial stability and regulation.

2020

• September 2020: The Common Market for Eastern and Southern Africa (COMESA) published a special report titled Role of Artificial Intelligence (AI) in Central Banking, discussing AI’s implications for central banks in the COMESA region.

2021

• July 2021: Researchers from the European Central Bank (ECB) released a paper titled Desiderata for Explainable AI in Statistical Production Systems of the European Central Bank, emphasizing the need for transparency and interpretability in AI applications within central banking.

• October 2021: A study titled Can an AI Agent Hit a Moving Target? explored the application of deep reinforcement learning algorithms to model economic agents’ behavior during monetary policy regime changes.

2022

• 2022: An article titled The Future of AI in Central Banking and Financial Services discussed the potential of AI to enhance regulatory efficiency and improve data analysis for monetary policy decisions.

2023

• January 2023: The Bank for International Settlements (BIS) published a bulletin titled Artificial Intelligence in Central Banking, highlighting the early adoption of machine learning techniques by central banks and the associated opportunities and challenges.

• July 2023: Piero Cipollone, a member of the ECB’s Executive Board, delivered a keynote speech titled Artificial Intelligence: A Central Bank’s View, discussing the implications of AI from a central banking perspective.

• September 2023: The Bank of Canada, under Governor Tiff Macklem, acknowledged that the adoption of AI by businesses could add to short-term inflationary pressures due to increased demand, while also noting the potential for AI to boost productivity in the long term.

2024

• October 2024: Sarah Breeden, Deputy Governor of the Bank of England, suggested that the rising use of generative AI by financial firms could be incorporated into the central bank’s annual stress tests, highlighting concerns about potential market manipulation and volatility.

• October 2024: The Bank of England launched a group of AI experts to study the risks and benefits of AI in the financial sector, aiming to understand how AI models used for trading could interact and impact financial stability.

2025

• January 2025: A Reuters article titled Maybe Fed Should Fret a Bit About an AI Wobble discussed how the U.S. Federal Reserve might need to monitor the impact of AI market disruptions on the broader economy, using the recent launch of the DeepSeek AI model as a case study.

• January 2025 : Bank for International Settlements. Governance of AI Adoption in Central Banks. BIS Representative Office for the Americas, January 2025.