By Seemanta Patnaik, Co-founder & CTO, SecurEyes
The potential of artificial intelligence (AI) is unlocking vast opportunities, particularly for central banks, as it brings considerable cost-saving and efficiency benefits besides providing cybersecurity. AI promises to enhance regulatory efficiency and bolster the foundational data used for making monetary policy decisions. That is why Central banks are rapidly deploying artificial intelligence (AI), driven by the promise of increased efficiency and cost reductions. According to findings from the Fintech Benchmarks 2023 report, central banks actively utilise various artificial intelligence (AI) and machine learning (ML) tools. Among a survey of 35 central banks, 51% reported the incorporation of AI/ML technologies. However, AI adoption is still slower in central banks than private financial institutions, which employ Blackrock’s AI-powered Aladdin, serving as the world’s top risk management engine to Robo-regulators in charge of ‘RegTech’ are ideal AI applications.
Machine learning (ML) has the capability to furnish comprehensive, instantaneous, and detailed information that complements the existing macroeconomic indicators. Its proficiency in efficiently analysing vast amounts of big data holds the potential to facilitate more informed monetary policy determinations. In response to the growing exploration and deployment of AI by countries and corporations, numerous financial authorities have formulated frameworks that articulate their expectations regarding the governance and utilisation of AI within financial institutions.
AI Engines Serving Central Banks
Athena, the novel natural language processing (NLP) artificial intelligence tool of the European Central Bank (ECB), plays a pivotal role in mitigating gaps in supervision and regulation. By enabling more than 1,000 supervisors to analyse an excess of 5 million documents within the Single Supervisory Mechanism (SSM), Athena processes diverse content, including news pieces, supervisory evaluations, and bank records.
This cutting-edge platform empowers users to seamlessly locate, extract, and compare information through a unified web-based interface. At the same time, Athena stands as a means for officials to identify and evaluate fresh risks at both macro and micro levels within the banking sector and assess them at an individual bank level.
To enhance the capacity of large language models, the ECB has established a cluster powered by graphics processing units dedicated to generating comprehensive insights at scale. Collaborative learning is another dimension of the ECB’s approach. It actively shares insights and practices related to AI model development and application with peer institutions. Beyond the confines of the Single Supervisory Mechanism, the ECB remains engaged in a continuous exchange of information on advancements in the field of NLP with counterparts like the Bank of England, the US Federal Reserve Board, and the Central Bank of Brazil. This ongoing interaction aims to harmonise best practices and explore avenues for deeper collaboration within the realm of AI and NLP, as stated by the European Central Bank.
AI for Cybersecurity in Banking
Most central banks that are part of the Financial Stability Benchmarks 2023 have reported a rise in cyberattacks targeting their countries’ financial sectors over the last two years. Of the 29 central banks surveyed, 24 (equivalent to 82.8%) have observed an escalation in electronic attacks directed at financial institutions. Furthermore, a significant portion of the surveyed central banks have expressed that they collaborate effectively with other agencies to address the challenges posed by cyber risks.
These figures suggest a rapid shift within the banking industry towards AI adoption aimed at enhancing cybersecurity measures. An illustration of this trend is evident in Danske Bank, Denmark’s largest financial institution, which has adopted an AI-driven fraud detection algorithm. By implementing this deep learning tool, the bank witnessed a substantial 50% boost in its ability to identify fraudulent activities, concurrently leading to a remarkable 60% reduction in false positive alerts. The AI-powered fraud detection system also streamlined critical decision-making processes while routing certain cases to human analysts for in-depth scrutiny.
Researchers at JPMorgan Chase have created an advanced warning system that employs AI and deep learning methodologies to detect malware, trojans, and phishing attempts. According to the researchers, it typically takes approximately 101 days for a trojan to infiltrate corporate networks. This early warning system furnishes substantial advance notice prior to the occurrence of an attack. Moreover, it promptly issues notifications to the bank’s cybersecurity team as cybercriminals gear up to dispatch malicious emails to staff members, intending to compromise the network.
Conclusion
The impact of AI within the banking sector is significant, prompting financial institutions to remain abreast of the latest developments. Within banking, AI-driven chatbots and virtual assistants play a pivotal role in enhancing customer service by delivering personalised assistance and adeptly handling common queries. Likewise, Robotic Process Automation (RPA) streamlines mundane manual tasks, such as data input and document processing, amplifying operational efficiency and reducing the occurrence of errors in banking procedures.
The application of AI is increasingly leveraged to scrutinize massive datasets, pinpoint patterns, and swiftly identify instances of fraud in real-time, thereby bolstering fraud prevention efforts. At the same time, utilising AI, banks can harness consumer data and preferences to offer tailored product recommendations, specialised promotions, and customised services, enriching the customer experience.
Given AI’s considerable advantages in the banking sector, FinTech enterprises are now channeling investments into AI for banking, particularly in intricate data analysis, forecasting market trends, and assessing risks. It contributes to more precise risk management strategies and informed decision-making processes.
However, alongside these opportunities, institutions encounter a range of challenges. Nevertheless, these technologies come with their own set of challenges. As per the Fintech Benchmarks 2023 report, these encompass concerns related to data quality (49%), the acquisition of pertinent skills (40%), ensuring algorithmic fairness (29%), contending with additional constraints (23%), safeguarding privacy (20%), and upholding ethical considerations (14%).
