What's new
Overview
This deep dive examines the latest developments in artificial intelligence, including emerging adoption patterns, key risk watch points, evolving policy and supervisory guidance, and research on the economic impact of AI.
What's new
Central Bank of the Republic of Uzbekistan rolls out nationwide AI training for regional banking executives
The Central Bank of the Republic of Uzbekistan and the Ministry of Digital Technologies have launched an artificial intelligence training programme for executive staff in the regional banking system, following a presidential instruction to expand digital technologies in banking. A two-day pilot course for Tashkent region bank branch executives covered AI applications in management, risk assessment, process automation, customer service, and related risks, compliance and governance, with similar courses to be rolled out across all regions.
European Central Bank working paper finds Q learning can amplify fund redemption fragility while large language models weaken coordination
The European Central Bank published a working paper analysing how different artificial intelligence architectures could affect financial stability via investor redemption behaviour in mutual fund “run” scenarios. The study finds that Q-learning investors coordinate easily but exhibit a systematic redemption bias under default risk, amplifying fragility, while large language model investors reason in expected-value terms and are less affected by default risk but display weaker coordination and lower predictability.
Securities & Exchange Board of India issues advisory on AI led vulnerability detection risks and creates cyber-suraksha.ai task force
The Securities & Exchange Board of India has warned securities market intermediaries about cyber risks from advanced AI tools and created a cyber-suraksha.ai task force for a coordinated response. The task force will assess AI-driven cybersecurity risks, develop a uniform mitigation strategy, facilitate threat intelligence sharing, and review third-party providers. SEBI’s advisory sets expectations for immediate and virtual patching, enhanced vulnerability assessments and SOC monitoring, stronger vendor and API controls, and longer-term plans for AI-based detection and mitigation.
Office of the Superintendent of Financial Institutions, Canadian Public Accountability Board and Canadian Securities Administrators highlight emerging risks to audit quality
Canadian financial and audit regulators highlighted technology, governance, fraud and disclosure risks affecting audit quality. They called for stronger oversight, risk-focused fraud defenses and more robust scrutiny of financial statement disclosures involving estimates, judgments and uncertainty.
French Financial Markets Authority research finds AI use in investing remains limited and mainly supports other information sources
The French Financial Markets Authority has published a research note on artificial intelligence in investment decisions by French adults, finding that AI remains a minority tool used mainly for information rather than as a standalone decision maker. Eleven percent of adults use AI before investing, with higher use among those under 35, and most users combine AI with personal research or professional advice. Respondents see potential for more tailored, lower-cost advice but also highlight risks of errors and reduced transparency, and the authority will monitor these trends.
U.S. Department of the Treasury hosts Liberty University summit on artificial intelligence energy and emerging technologies
The U.S. Department of the Treasury hosted a summit with Liberty University on artificial intelligence, energy and emerging technologies, focusing on their impact on the economy and public sector delivery, without announcing new policy measures. Four panels addressed frontier AI, energy, workforce productivity and financial literacy, with participation from senior U.S. government officials and industry representatives.
Federal Reserve Board's Bowman outlines AI supervision after excluding generative and agentic AI from model risk guidance
The Federal Reserve Board published remarks by Vice Chair for Supervision Michelle Bowman outlining a supervisory approach to banks’ use of artificial intelligence that focuses on material financial risk while avoiding unnecessary impediments to innovation. She noted that the Federal Reserve, Office of the Comptroller of the Currency, and Federal Deposit Insurance Corporation have amended model risk management guidance to exclude generative and agentic AI, with these tools instead covered by broader governance and risk-management expectations. Bowman also said the Financial Stability Board’s Standing Committee on Supervisory and Regulatory Cooperation is preparing a consultation report on sound practices for AI adoption, including potential needs for greater international consistency in expectations for cybersecurity and critical infrastructure.
European Insurance and Occupational Pensions Authority flags geopolitical tensions and market risks in April 2026 IORP dashboard
The European Insurance and Occupational Pensions Authority has published its April 2026 risk dashboard for institutions for occupational retirement provision, highlighting elevated geopolitical tensions, higher inflation projections and market risks as key concerns. While bond spreads have widened and equity and corporate market volatility has increased, liquidity and funding risks are contained and the financial position of defined benefit IORPs has strengthened. The dashboard covers defined contribution and defined benefit schemes in the European Economic Area, based on Q4 2025 regulatory data from 625 IORPs and market data up to end-March 2026.
Australian Prudential Regulation Authority calls for step change in AI risk management after review finds governance and cyber controls lagging adoption
The Australian Prudential Regulation Authority has issued a letter to regulated entities outlining findings from a 2025 supervisory review of artificial intelligence use, warning that governance, risk management, assurance and operational resilience are not keeping pace with rapidly expanding deployments. The review highlights board-level technical literacy gaps, lagging information security controls, fragmented assurance, concentration risks in AI suppliers and weak contingency planning, and sets expectations for stronger oversight, cyber hygiene, human involvement in high-risk decisions and integrated assurance. APRA is finalising a supervisory plan and signalled it will escalate enforcement where entities fail to manage AI risks proportionately.
South Korea Financial Services Commission outlines targeted fintech support upgrade and stablecoin and data reform plans
South Korea’s Financial Services Commission outlined plans to overhaul its fintech support framework to focus more on artificial intelligence, data, regional enterprises and young entrepreneurs, alongside regulatory changes to establish a legal basis for stablecoins and ease data-use rules. The initiative will introduce performance-based incentives, strengthen AI and overseas expansion programs, create regional public-private governance networks, and establish an integrated one-stop fintech support information platform.
Canada's Office of the Superintendent of Financial Institutions outlines priorities on credit non-bank and liquidity risks and signals targeted capital and liquidity calibration
The Office of the Superintendent of Financial Institutions is sharpening its supervisory and policy agenda around risks in its Annual Risk Outlook, including real estate secured lending, non-bank financial institutions, and liquidity and funding risks, supported by a new Credit Risk Management Guideline consultation, liquidity framework refinements, and earlier supervisory engagement with boards and senior management. OSFI has streamlined its policy suite, signalled willingness to recalibrate capital and liquidity requirements, and continues to benchmark Canadian capital rules internationally. On climate, tokenization, stablecoins, and artificial intelligence, OSFI is pursuing a measured, prudentially focused approach that emphasises governance, accountability, and keeping activity within the regulated perimeter.
Japan Financial Services Agency establishes finance sector cybersecurity working group to address AI driven threats
The Japan Financial Services Agency has established a public-private working group on financial sector cybersecurity, described as a Japanese version of “Project Glasswing”, to coordinate responses to AI-driven cyber risks. The initiative aims to accelerate information sharing on vulnerabilities, shorten the time from identification to patch application, and strengthen preparedness for incidents in the highly interconnected, real-time financial system, with potential expansion to include IT firms, online finance participants, and engagement with international counterparts through G7 and G20 forums.
National Bank of Moldova deputy governor sets out AI resilience and accountability risks at the IMF New Economy Forum
The National Bank of Moldova reported that Deputy Governor Mihnea Constantinescu spoke at the IMF’s “New Economy Forum: AI and the Resilience Gap”, warning that uneven AI adoption and concentration in data infrastructure and frontier models can create systemic dependencies and widen resilience gaps between countries. He stressed that effective AI integration depends on public and private sector digitalisation and highlighted governance risks in hybrid human-algorithm decision chains, where autonomous AI agents may undermine accountability.
International Monetary Fund publishes note on how agentic AI could change payment authorization and settlement
The International Monetary Fund has published a note on how “agentic” artificial intelligence, which can act autonomously on users’ behalf, could reshape payment systems by shifting transactions to agent-mediated decisions. Using a three-layer framework of intent, authorization, and settlement, it assesses potential benefits and risks for authorization, liquidity, settlement, compliance, and resilience, highlighting tensions between probabilistic AI behaviour and deterministic payment infrastructure requirements, and risks around traceability, opacity, systemic effects, cybersecurity, and legal uncertainty.
United Arab Emirates Cabinet adopts agentic AI framework targeting 50% of government services within two years and approves new digital records and data sharing policies
The UAE Cabinet approved a new federal framework to deploy agentic artificial intelligence across government operations, targeting the transition of 50% of government sectors, services and operations to autonomous AI models within two years, supported by unified standards and data policies. Implementation will be overseen by H.H. Sheikh Mansour bin Zayed Al Nahyan and a dedicated taskforce, with ministries assessed on AI adoption and complementary measures including a federal Code for Government Services, a Zero Bureaucracy reference model, and new policies on digital records and data sharing.
U.S. Senate Committee on Banking, Housing and Urban Affairs Ranking Member Elizabeth Warren warns of a debt-fuelled AI bubble and urges pre-emptive structural reforms
The U.S. Senate Committee on Banking, Housing and Urban Affairs published remarks by Ranking Member Senator Elizabeth Warren warning that a debt-fuelled boom in artificial intelligence investment could pose systemic financial stability risks. She highlighted heavy reliance on private credit and complex debt structures, growing interconnectedness and opacity, and questioned whether AI revenues can scale sufficiently to service projected borrowing. Warren reiterated calls for structural financial reforms, a new digital regulator, and stronger accountability measures to mitigate potential AI-related financial shocks and avoid future bailouts.
Luxembourg's Ministry of Finance advisory board publishes first recommendations on integrating AI in finance
The Advisory Board on AI in Finance of Luxembourg’s Ministry of Finance has issued its first recommendations on integrating artificial intelligence in the financial sector, reflecting AI’s growing impact on financial services and competitiveness. The forward-looking guidance outlines strategic perspectives on AI-related opportunities and risks to support Luxembourg’s position as an international financial centre.
Dubai International Financial Centre announces plan to become the world’s first AI-Native financial centre
The Dubai International Financial Centre plans to become the world’s first AI-native financial centre, embedding artificial intelligence across its legal and regulatory frameworks, business environment, talent development, infrastructure and urban design. Building on its five-year AI strategy and AI-specific data governance under the DIFC Data Protection Law, the programme will establish ethics and governance frameworks for AI agents and robotics, provide firms with advanced AI tools, and develop an AI Campus, with DIFC estimating USD 3.5bn in economic benefits and 25,000 new jobs.
National Bank of Moldova governor outlines internal AI governance and planned rules for AI cloud and eKYC
National Bank of Moldova Governor Anca Dragu outlined the central bank’s gradual approach to artificial intelligence adoption and oversight, highlighting current financial sector uses in credit risk assessment, liquidity management, forecasting and fraud detection. She detailed the bank’s internal AI governance framework, AI-based assistant, automation and SupTech data capabilities, and noted forthcoming regulations on AI use, cloud services and electronic customer identification, underpinned by a “human-in-the-loop” principle to mitigate cybersecurity, data protection and third-party risks.
Bank for International Settlements publishes dataset on 1,246 AI-producing firms and finds the United States and China are the largest hubs
The Bank for International Settlements published a working paper and cross-country dataset mapping 1,246 AI-producing firms across 32 economies into five AI supply chain layers, highlighting global concentration, specialisation and investment patterns relevant to AI industrial and sovereign strategies. Firm activity is heavily concentrated in the United States and China, with significant economic footprint in Chinese Taipei and Korea, and shows strong home bias in investment and a high share of deals targeting AI applications. The publicly available dataset is based on PitchBook data, uses large language model-assisted classification with manual verification, and covers firms with valuations above USD 500 million.
UK Parliament Treasury Committee publishes regulators’ AI responses as Bank of England plans AI-agent trading tests
The UK Parliament Treasury Committee has published responses from the Bank of England, HM Treasury and the Financial Conduct Authority to its report on artificial intelligence in financial services, including confirmation that the Bank plans to test AI agents in financial trading markets and assess correlated “herding” risks. HM Treasury did not commit to bringing major AI and cloud providers into the Critical Third Parties Regime before the end of 2026. The Financial Conduct Authority plans to share AI best practice examples with firms, and the Committee will continue to monitor progress and potential disruption from outages at major providers.
Spanish Securities Commission publishes study warning that unsupervised large language models can produce faulty stock investment recommendations
The Spanish Securities Commission (CNMV) has published a research paper finding that advanced large language models used for stock investment predictions pose significant operational risks and potential investor losses without human oversight. Comparing outputs from ChatGPT, Gemini, DeepSeek and Perplexity, the CNMV highlights recurring reasoning failures, especially with simple, unstructured queries, and concludes that robust governance, systematic human validation and grounding models in official, regulated data are essential to improve reliability and reduce errors.
Dutch Authority for the Financial Markets publishes report on AI in capital markets and stresses human oversight and clear accountability
The Dutch Authority for the Financial Markets has published a sector report on how artificial intelligence is transforming capital markets trading, highlighting efficiency gains alongside heightened integrity and vulnerability risks. It flags risks from self-learning trading systems, information-based manipulation and multi-agent technology, and stresses the need for clear human responsibility, robust controls and oversight. The authority will strengthen its supervision and expertise, translate emerging risks into supervisory measures, consider further regulatory changes or interpretations, and continue engagement with market participants and other supervisors.
Central Bank of Cuba reviews artificial intelligence strengths and risks and sets out priorities for responsible use
The Central Bank of Cuba published an international overview on how artificial intelligence is reshaping sectors including finance, highlighting both its benefits and ethical, social and economic risks. The article outlines key applications in healthcare, finance and education, and recommends explainable and auditable models, legal and ethical frameworks, workforce reskilling, equitable access and stronger safeguards against cyberattacks and data manipulation.
State Bank of Vietnam supports an ASEAN common baseline framework for AI governance and stronger regional cooperation to counter digital fraud
The State Bank of Vietnam reported on Deputy Governor Nguyen Ngoc Canh’s participation in an ASEAN dialogue on a common baseline framework for artificial intelligence governance in finance and an ecosystem-based approach to rising fraud and scams. Canh backed a voluntary, risk-based, purpose-specific regional framework prioritizing operational guidance for banking use cases such as fraud prevention, anti-money laundering, countering the financing of terrorism, customer service and decision support. Participants also urged stronger coordination and information sharing among banks, regulators and other stakeholders to combat increasingly digital, cross-border fraud.