The Pulseline News Desk
Central banks must adopt artificial intelligence (AI) under strict governance and human oversight to prevent emerging technologies from undermining public confidence in financial systems, International Monetary Fund (IMF) Chief Information Officer Shirin Hamid has warned.
Delivering the 76th Anniversary Oration of the Central Bank of Sri Lanka (CBSL) in Colombo, Hamid has cautioned that while AI offers significant opportunities to improve financial operations, its rapid adoption also exposes institutions to increasingly sophisticated cyber threats, fraud and data-related risks.
She has identified three major threats confronting financial institutions, which are impersonation through cloned voices and deepfake videos, autonomous AI systems capable of launching cyberattacks at machine speed, and indirect attacks designed to manipulate AI assistants operating within institutional networks.
Hamid has warned that technology now allows a person’s voice to be cloned using as little as three seconds of recorded audio, raising concerns about the potential for sophisticated impersonation and financial fraud.
She has also highlighted the risks associated with relying on generative AI for economic analysis, particularly when inaccurate information is presented with apparent confidence.
According to Hamid, IMF testing has found that commercial generative AI tools, including ChatGPT, had returned inaccurate gross domestic product figures in as many as two-thirds of the cases examined.
Such errors could have serious consequences for institutions whose credibility depends on the accuracy of economic data, she has said.
In response to these challenges, the IMF has developed its own system, StatGPT, designed to retrieve information directly from verified datasets rather than relying on AI models to generate potentially inaccurate figures.
The example, Hamid has suggested, demonstrated why financial institutions must prioritise the reliability and verifiability of information when integrating AI into their operations.
Trust must underpin AI adoption
Hamid has argued that the effectiveness of an AI-enabled central bank should not be measured by the number of AI systems it deploys, but by its ability to ensure that those systems operate securely, ethically and transparently. “The intelligent central bank is not the bank with the most AI. It is the bank whose AI can be trusted and can ensure that any AI that touches the financial system is explainable, governed, secure, ethical, and human-supervised,” she has said.
She has identified four essential pillars for the responsible adoption of AI by monetary authorities: dependable data, trustworthy systems, operational resilience and adaptive governance.
These principles, she has said, must guide technological transformation across central banks, rather than allowing AI adoption to become an isolated information technology initiative.
Hamid has stressed that human judgment remained indispensable, particularly when AI systems were used to support decisions affecting financial stability, monetary policy and public confidence.
“For every AI-influenced decision, we need a named human who is answerable,” she has said, adding that institutions should seek to strengthen human judgment before granting greater autonomy to AI systems.
Implications for financial stability
The warnings are particularly relevant as financial institutions worldwide explore AI applications in data analysis, risk management, fraud detection and customer services.
While such technologies can improve efficiency and help institutions process large volumes of information, vulnerabilities in their design or implementation could expose financial systems to new forms of operational and cyber risk.
For central banks, the challenge is therefore not simply to keep pace with technological advances, but to ensure that innovation does not compromise the integrity of their operations or the public’s trust in financial institutions.
Hamid has concluded that although financial technologies would continue to evolve, the fundamental responsibility of an intelligent central bank remained unchanged: protecting public confidence in the financial system, from traditional banking operations to AI-driven decision-making.
Leave a comment