Banki szakember
Every story that matters to the banking professional, with one sentence on what it concretely means for them. 442 stories, newest first.
November 2026
88-year-old loses $450,000 in deepfake scam; bank will not reimburse under Canadian rulesMulti-step confirmation is advisable for large transfers by older customers.ECB may review banks' data center lending over AI riskThe data center lending portfolio should be reviewed in anticipation of an inspection.Gránit Bank's digital banking and AI expertise to be built into university teachingThrough an education partnership, the bank is building a future workforce with AI skills.Three charged with stealing $129 million from companies by impersonating CEOs with AIFor large, urgent transfers by corporate clients, multi-step confirmation is advisable, and payments to crypto should be monitored separately.Visa test: AI shopping agents failed 42.8% of runsWith agents initiating card transactions, the high failure rate means banks should prepare for more disputes and refund cases.Anthropic adds personal finance feature to Claude, with bank account connectionGeneral-purpose AI assistants compete with banks' own financial assistants.JPMorgan Chase partners with OpenAI on AI for cybersecurity and engineeringCoding and cybersecurity are emerging as top internal AI use cases at large banks.Gránit Bank: customers increasingly open to banking AI for managing spendingCustomers' spending-analysis questions show the main use case of a bank's AI assistant.Hedge funds shed AI debt as $500 million data center loan package heads to auctionThe sale of data center loans on the secondary market shows that the pricing of AI infrastructure exposure is changing.Kakao Bank's AI call bot cuts fraud transfers and raises cancellation rate by 54%An AI call before a transfer can serve as a working fraud-prevention control.Visa expert: bank AI projects rarely fail because of the modelFactors outside the model, such as embedding in processes and the organisation, decide whether an AI project succeeds.Indian authorities warn of AI-powered attacks on financial institutions with unauthorised transfersBased on the official warnings, transaction monitoring and incident response must be scaled for machine-speed attacks.OCC's first supervisory guidance on bank AI use rates simple payment agents as high riskInventories, risk classifications and control documentation for AI systems should be prepared for examinations, especially for agents that make payments.Agent identity: LexisNexis, Mastercard and Trulioo build Know Your Agent toolsFor transactions initiated on a customer's behalf by an agent, banks must decide which identification signals they will accept.One in five young Australians turns to AI for financial advice, and many act without checkingClients can compare an adviser's role with AI, so a transparent explanation is an advantage.Visa research: how banks in Central and Eastern Europe use AIData on regional banks' AI use provides a benchmark for one's own rollout plan.Visa presents a solution for AI agents' access to card dataA network-level solution is emerging for securely issuing card data access to agents.Hong Kong becomes first banking regulator to require identity credentials for AI agentsThe agent's identifier must be recorded in the onboarding and risk-management process.MAS Future of Finance Institute starts operations in Singapore to speed up AI adoptionIn the Singapore market, AI adoption may speed up with supervisory support.HDFC and Kotak: AI agent safeguards have not kept pace with data access, The Ken reportsAgents' data access should be limited with least-privilege permissions and logging.Mashreq adds Clari5 generative AI to strengthen fraud risk managementIn fraud detection, generative AI is being brought in to support analysts' work.JPMorgan, Amazon and Bank of America restrict external AI agents' accessAgent access needs a dedicated policy and customer communication; general terms of use are not enough.Goldman Sachs limits advanced AI coding tools to engineersIn non-engineering roles, use of advanced AI coding tools may be banned or require approval at some large banks.FCA sets out good and poor practice for AI use in financial servicesAI use in financial crime prevention should be checked for the shortcomings the FCA has flagged.ECB lists AI stock valuations among financial stability risksRisk managers have reason to review AI-related exposures.Barclays and NatWest job cuts seen as first visible AI effect on banking employmentHiring may slow for entry-level and customer service roles, so it is worth building skills that complement AI.Survey: 65% of US consumers would blame the AI company if an AI agent commits fraudGiven the customer trust gap, agent features should be introduced in stages, with transparent liability rules.OTP: fraudsters take an average of HUF 1.4 million per victim, and AI heightens the threatIn customer communication, customers must be prepared for AI-perfected phishing.Authorities and media warn of AI voice-cloning scamsInforming customers about such scams and warning them before a transfer can reduce losses.Philippine central bank: central banks must weigh AI risks and opportunitiesASEAN banks can expect guidelines on the responsible use of AI to be drawn up.HSBC to open AI institute in Hong Kong following Singapore launchLarge banks are building regional AI research hubs to assess new technologies.Analyses: AI agents could shake bank deposits and payment authenticationDeposit-retention and authentication strategies should be prepared for agent-driven money movement.Colorado credit union and two Philippine firms roll out AI scoring for lendingWith AI-based loan pricing and scoring, explainability and bias testing must be ensured.Bank of England says AI assurance practices have gapsThe validation and verification of AI models is not yet mature enough, so an internal testing framework needs to be built.Morgan Stanley advisors have built nearly 50,000 AI agentsAgents built by advisors need a regulated framework, so the control model should be worked out in advance.Hungarian-founded Zocks named to CB Insights Fintech 100 listAdviser assistants speed up the administration of investment advice.Only one in five people in Hong Kong trust banks to stop deepfake scamsProtection against deepfake scams should also be communicated visibly to customers.
Oct 2026
ECB's Lane: AI is a significant driver of euro-area growth and lendingAI-related investment and credit demand is one of the drivers of euro-area loan portfolios.Agentic AI is reaching production in banks faster than supervisory guidance can keep upGuardrails and supervisory expectations should be built into deployments from the design stage.Barclays seeks credit risk insurance for AI data center financingTransferring risk on AI infrastructure loans could open up new capacity and financing room.Bank of America and Capital One use smaller AI models to cut costsIn AI deployments, balancing model size against cost is an operational decision.Canada's big banks, including RBC, BMO and TD, block third-party AI agentsDetecting and blocking agent access, and communicating with customers, are new tasks for digital channels.Moody's and BIS flag private credit and financial stability risks as AI-related debt growsWhen assessing AI clients, lenders should examine business and financial viability; an AI label is not enough.AI lifts bank productivity in Asia but does little for revenueThe rollout of autonomous agents must be planned together with controls.Conquest's verifiable AI advice engine joins Anthropic's Claude for Financial Advisors, which adds tax-aware analysisFor AI built into advisory workflows, verifiability and the handling of tax effects are the main considerations.Major U.S. banks block AI agents that log in with customers' credentialsBanks already struggle to tell whether the account owner or an AI agent is carrying out an action, so identifying agent traffic is an urgent task.Bank of England and FCA publish joint report on AI in UK financial servicesWhen adopting agentic AI, documenting vendor dependence and governance may become an expectation.Barclays to cut about 5,000 jobs as it leans on AIAutomatable back- and middle-office roles are the targets of the cuts, so it is worth retraining toward supervisory and exception-handling skills.Goldman Sachs and JPMorgan take leading roles in Anthropic's IPOIn investment and private banking, retail allocation and the handling of AI exposure bring new client questions.Fraudsters use AI voice imitation to steal huge sum from Italian bankVoice-based orders and executive instructions require independent, multichannel confirmation.FCA builds agentic finance rules without introducing AI-specific regulationTesting and quality-assurance procedures for agentic systems need to be tightened.Deutsche Bank uses AI to cut some corporate loan decisions from weeks to minutesIn credit assessment, the role of AI and the rules on human approval are part of the competition on turnaround time.Shopify adopts Visa, Mastercard and American Express agentic payment technology for AI shopping agentsIssuers need to prepare to authenticate and fraud-screen the tokenized cards that agents use.McKinsey: agentic AI could cut banks' transaction monitoring workload by up to halfThe work of transaction monitoring analysts is shifting towards reviewing agents' outputs.Banks hand more decisions to AI as governance lags behindThe balance between risk and innovation must also be applied when launching products.Central and Eastern European banks, including Hungarian ones, catch up on AI useAI applications in lending and cybersecurity are the main areas where regional banks have ground to make up.One in three Europeans would rather discuss sensitive financial matters with AIOn sensitive topics such as payment difficulties, an AI channel can be customer-friendly if a handover to a human is available.European payments firms form council on agentic commerceTeams responsible for payment products need to follow the council's recommendations.Dutch police arrest two men over AI-call and NFC relay bank fraudCombined patterns of NFC relay and telephone deception need dedicated detection rules.48% of Hungarians use AI for finance or insurance questions but choose humans for big decisionsAlongside AI-supported information, customers should also be offered access to human advice.Starling Bank's AI system fully investigates a suspicious transaction in 17 secondsFraud investigation turnaround time has become measurable and competitive, and can be used as a benchmark.Hungary's central bank examines AI strategies at banks: plenty of risk, few peopleAn AI strategy should also document the level of expertise and human oversight.K&H turns to conversational banking and presents its Kate assistantA conversational interface is reshaping the division of work between digital channels and customer service.Hungarian trade press: AI is rewriting the roles of finance professionals and executivesModelling and data skills are becoming more valuable.Mastercard's Hungarian expert: by 2030 the AI agent pays and the token verifiesIn payment and card products, it is worth preparing token-based agentic authentication.Erste reports nearly 21,000 fraud attempts against its customers in six monthsFraud-prevention communication should also target farmers and small business owners.US regulators expect banks to account for AI use; SEC to focus on AI in examsDocumentation on AI system vendors and security should be put in order.Banks call for stronger intelligence sharing to fight AI fraudFraud indicators must be shared between institutions, which requires data-protection frameworks.CyberCartel reportedly used Claude and Grok AI agents to automate attacks on Latin American banksAutomating attacks speeds them up, so detection and response must be made faster as well.Barclays to cut 5,000 jobs, about 5% of its workforce, as it steps up AI adoptionSupport and operations roles are the most exposed, so upskilling and retraining for these roles is urgent.48% of Hungarians use AI for financial questions but seek human advice on big decisionsCustomers increasingly arrive with preliminary information obtained from AI, which has to be handled in client service.European financial firms move to open AI models run on their own infrastructureOpen models run on in-house infrastructure give better data control, but the bank is responsible for operating them.JPMorgan's private bank uses AI to make advisers more efficient; Goldman offers small businesses an AI packageAdvisory work is increasingly AI-assisted, and compliance with security rules is a key issue.Bank of Ireland accelerates AI in fraud detection, with 1 billion checks and €9.7m savedAI-based transaction checks can deliver measurable savings at high volume.Banks and central banks examine AI's impact on employmentLabour-market risk is greatest for routine back-office tasks, so it is worth building skills ahead of time.K&H's Kate assistant interprets customer requests in conversation and can initiate transfersConversation-based transactions require rules for authentication and error handling.Bank AI projects: OpenAI on COBOL migration, Danske Bank API agents, BNK group-wide platformAI-assisted migration of legacy core banking code and agent-API integration bring both operational risk and opportunity.Veryfi launches AI suite to screen checks, invoices and receipts for document fraudReal-time API-based screening against check and document fraud can be built in.Bank of Japan, MAS and China's central bank focus on AI risks and strategyScenarios of AI-driven cyberattacks and market shock may be included in risk stress tests.Hungarian banks and payment players on AI-agent shopping: OTP, SimplePay, WoltIn agent-based shopping, the bank's role and its proximity to the customer are a strategic decision.Fitch sees moderate AI disruption risk for wealth managers and retail banks; BCG says finance is most exposedIn retail banking, AI-based advice creates competitive pressure.EOS survey: 46% of European consumers would discuss sensitive banking matters with AIA significant share of customers are willing to raise sensitive topics through an AI channel, so data protection and handover to a human agent matter.US big banks block third-party AI agents from accessing customer accountsRecognising agent traffic and tightening customer authentication rules is a direct operational task.Mastercard opens early access to Agent Pay for Machines for AI agentsIssuers must prepare for agent tokens and for authorisation rules covering agent-initiated transactions.FCA sets out good and poor AI practice; firms overestimate model accuracyClaims about model accuracy must be backed by documented testing.Credit markets look for weak links in AI debt as credit spreads hit a 17-month highExposure to AI data-centre financing should be reviewed against concentration limits.Erste Bank customers targeted nearly 21,000 times in six months, with losses of HUF 7 billionFraud prevention must also be ready for transactions that are carried out by the victim and look legitimate.World launches World ID suite for banks to counter AI-driven fraud and deepfakesCustomer identification and video-call authentication can be supplemented with proof of human presence.AI voice-cloning and fake-retailer scams target retail customersTargeted warnings and delays can be applied to urgent transfers initiated by customers.Stablecoins compete for AI-agent payments, with USDC in the leadBanks should assess which payment flows could move off bank rails through agents and stablecoins.Deepfake executive scams: $95 million loss, FBI data, INTERPOL and KeyBank warningsSignals of executive-impersonation fraud against customers must be built into transaction monitoring.Morgan Stanley launches UK savings accounts as US banks fear AI agents could drain cheap depositsAI agents could weaken deposit loyalty, so banks should prepare for rate competition.FIS and Anthropic's Claude-based money laundering investigation agent is available to any U.S. bankInvestigators' work shifts toward reviewing the material the agent assembles.HKMA launches "ART" framework for responsible AI adoption in banking; banks test deepfake detectionThe framework for responsible AI adoption can serve as a reference for internal governance.Huntington Bank launches AI business-startup platform; APG Federal Credit Union goes live with Scienaptic AIAI-supported credit-assessment and small-business onboarding processes are available to smaller institutions too.G7 to test banking system defences with AI-assisted cyberattack simulationsExpect your incident response and business continuity plans to be tested against AI-driven attack scenarios.K&H press briefing covers AI portfolio management and ETFsETFs and AI support now appear together in the retail investment offering.OTP Bank launches film series and expert warning against AI-driven fraudCustomer education and communication against social engineering are a key element alongside technical defences.FCA to review suitability tests for complex products as AI simplifies questionnairesWhen selling complex products, it is worth building in additional checks.Anthropic's first credit rating revealed as BB+, with $60 billion financing loomingWhen lending to AI companies, the rating is one input to risk pricing.Colony Bank uses generative AI to cut days from SBA loan approvalsDocument processing in credit underwriting is one area where AI pays back quickly.OpenAI discloses four more incidents involving financial, medical and government dataIncident reports from AI providers should be built into third-party risk monitoring.In Slovakia, AI writes bank complaints, winning customers over half a million eurosIn complaint handling, expect AI-written submissions that are legally more precise.Deutsche Bank launches its own AI platform for its investment businessLarge banks are building their own AI platforms into investment processes.FSB's final report sets out 12 sound practices for responsible AI adoptionCompare the human-oversight points in your agentic AI processes against the 12 practices.Erste says it saved HUF 6 billion from fraud, with 21,000 attemptsThe figures show that transaction monitoring and customer education together deliver measurable savings.Ping An Digital Bank reaches 96% deepfake detection accuracy in HKMA GenAI sandboxThe sandbox result is a benchmark for the accuracy of deepfake-detection systems.Hungarian experts: scammers use AI and psychological pressure to drain bank accountsIn customer communication, it is worth repeating call-verification and call-back procedures continuously.Report finds banks may face liability over use of AI chatbotsA bank may be liable for the answers of a chatbot on a customer-facing channel, so supervise and log them.Italian bank loses $108M to AI messaging fraudIn the approval chain for large transfers, check controls against impersonation by messaging.EOS Banking AI Barometer: AI lives up to expectations at banksThe barometer is a benchmark for comparing Hungarian and regional AI use in banking.FICO survey: 77% of Asia-Pacific banking leaders want purpose-built AIFraud detection is the primary banking use case for AI, with data privacy the main brake.CarbonAI raises $12M and Matters $5M for AI lending and AI-native bankingIn commercial real estate lending, expect new AI tools.Apate.ai uses AI agents to trap scammers, learning from 2.5 million conversationsData from scammer conversations can be used to train fraud-detection models.FCA to automate supervision with AI, with no new rules for chatbotsCompliance is judged under existing rules rather than new chatbot rules, so it is worth documenting internal controls.American Express publishes its standard for AI agent paymentsPre-payment agent verification offers a model for reshaping fraud screening.Coinbase test: newer AI models missed more payment fraud casesDo not assume a newer model is more accurate in fraud screening: retest whenever you switch models.Google adds Gemini-based scam protection to Android and SingaporePlatform-level protection may reduce the number of scams, but it does not replace the bank's own controls.Report: deepfake fraud market reaches $40 billionThe scale justifies building deepfake detection into authentication processes.Bank regulators in the US and Europe tighten scrutiny of AI risksCompile an inventory of AI use, including components of Chinese origin, and have a documented shutdown mechanism for every autonomous agent.IMF stability report flags AI market correction, rising yields and cyber riskCheck AI-related credit and market exposure.SBS and NVIDIA launch sovereign AI foundation for core banking securityA solution with sovereign data handling is available for core banking AI and is worth examining.MBH's BUPA launches AI pre-screening to help Hungarian SMEs find fundingPre-screening automates the early stage of the SME lending process.JPMorgan blocks outside AI agents from customer accounts; Schwab also blocksIdentifying and blocking agent traffic, and handling customer consent, should be written into policy.AI financing tightens as issuance falls and default protection gets costlierTreat AI-infrastructure customers separately within credit risk limits.Big banks acknowledge AI financing risks but keep lending to the boomIn syndicating AI loans, the large banks' caution can serve as a benchmark for pricing and risk-taking decisions.Banks seek customer trust against AI voice cloning as SBI warns and call protection spreadsCustomer calls must be made verifiable and recognisable.Report: AI could free up 40% of banks' middle and back-office capacity by 2030Entry-level tasks are shifting toward supervising AI agents, so it is worth building skills for that.AI agents challenge payment intent verification and fraud liabilityDevelop intent and consent checks for transactions made by agents acting on a customer's behalf.AI-driven cyberattacks hit four South Korean banks; financial regulator orders bank-wide checksReview access and vulnerability-management processes on the assumption that attackers are AI-assisted.Singapore regulator sets AI rules for banks and insurersAt banks with a presence in Singapore, check alignment with the new AI rules.Superbank turns to AI to look beyond credit scoresFor AI credit scoring based on alternative data, document the model's explainability.Visa survey: consumers trust AI to shop, but not to payExpect agent-initiated payments to appear during the holiday season.Gránit Bank survey: customers open to AI-based financial helpCustomer demand provides a basis for introducing AI assistant features.ANZ chief warns of AI risks and possible job cuts as banks compete for AI talentRegulatory and risk-management AI expertise is in demand, so further training offers a competitive edge.Bank earnings hinge on Anthropic IPO as SpaceX windfall fadesAI companies' stock market listings directly affect the earnings of the investment banking business.Goldman Sachs expects junior bankers to manage AI agentsJunior roles are shifting from analyst work towards supervising agents and checking their outputs.Deepfake attacks target corporate finance; one in five finance leaders fear impersonationUnusual instructions received in a customer's name require multi-step verification.Incognia launches detection service for fraud committed by AI agentsIn digital-channel risk scoring, it is worth examining how bot and agent signals are handled.Community financial institutions rank fraud detection and document handling as top AI prioritiesSmaller institutions are focusing on fraud detection and document processing, so that is where benchmarking makes sense.HSBC to cut half of UK wealth management and specialist roles amid AI pushAdvisory and specialist roles are the most exposed, so it makes sense to move towards skills that complement AI.Singapore's MAS issues risk management guidelines on banks' use of AIInventory, oversight and ownership of AI models require an internal process.Ray Dalio and a London strategist warn of AI bubble burst; former World Bank governor sees no bubbleStress tests of AI-linked collateral and loans should be broadened.JPMorgan tops global AI banking index for fifth straight year; UBS only European bank in top 10Compare your own bank's AI maturity with the leaders against the ranking's criteria, and assess the gap.Dimon warns AI could raise cyber risk tenfold, citing an Anthropic modelCyber-risk models need to be extended with scenarios involving AI-capable attackers.Plaid launches LendScore 2 and pushes AI deeper into lending and fraud checksA new scoring model is available for lending decisions; it must be validated and checked for explainability.Norway's DNB to cut about 400 jobs as AI takes over manual tasksReplacing manual back-office tasks is increasingly a headcount-planning question for banks.SEON introduces a Claude-based AI assistant for fraud detectionAn AI assistant can be available to expand investigation capacity, but human decision-making is still required.South Korean banks breached with AI tool; CrowdStrike points to likely China-based attackerCredential-stuffing defences at customer login (MFA, rate limiting, anomaly monitoring) need to be reviewed now.Mastercard expands Agent Pay as card networks and Stripe build competing agent protocolsCard issuers must work out how to identify and authenticate agent-initiated transactions.AI voice clone and deepfake used to trick Italian bank into wiring €95 millionFor large transfers, a voice or messaging request must be confirmed through an independent channel.AI agents could run on stablecoins: BlackRock, Coinbase and new issuers make movesMachine-to-machine payments are a new type of traffic that also affects banks' custody and settlement offerings.Agentic payments: BIS, BCG and industry analyses on controls and risksTraditional fraud detection can mislead on agent-initiated transactions, so new signals need to be introduced.JPMorgan sells Volta's $5bn AI loan at 11% yield and leads $90bn Alphabet financing as banks weigh AI debtSelling and syndicating AI infrastructure loans is a new channel for transferring risk.FICO: AI arms race is reshaping the fraud threat across EMEAFraud prevention should prepare for higher-friction, continuous authentication.K&H and Mastercard extend partnership to 2030, adding AI-based payments to joint developmentAI-based payments will be a priority in K&H's card product development in the coming years.First test of AI-based payments in Hungary completed; OTP and SimplePay respondHungarian banks and payment providers must decide how to handle purchases made by AI agents.15% of Hungarians would already give their bank card details to AIHanding card details to AI raises fraud and liability questions, so banks need to address them in customer communication.Four Magyar Nemzeti Bank presentations on PSR/PSD3 payment rules and AIPreparation for the EU changes affecting payment services can begin.Voice-cloning scams: advice for consumers and new AI-driven fraud methodsIn customer communications, stress the details a bank will never ask for in a conversation.Hungary's cybersecurity institute and experts warn of AI-made financial fraud and cyberattacksBased on the cybersecurity warnings, firewall updates and customer alerts must be carried out in good time.Australian bank union says AI chatbots are increasing banking workloads and affecting customersCases the chatbots fail to resolve flow back to human staff, so the workload needs to be measured.Dubai Police: three seconds of audio is enough to clone a voice, warning of financial fraudThe reliability of voice-based customer authentication needs to be reassessed.TD Bank survey: 18% of Americans would let AI make financial decisions on its ownFor agent-based services, autonomous decision-making authority needs to be limited and customer consent handled explicitly.AI tools suspected in hack of South Korea's Shinhan BankBank defences must account for intrusions accelerated by AI tools.Nearly half of UK financial leaders use agentic AI; many firms have limited controlAgentic workflows need a traceable decision log and human oversight points.AI is redefining Wall Street jobs, with demand for one new skill up 1,721%Acquiring new AI-related skills is becoming increasingly important for labour-market position.Hungary's central bank says AI increases European banks' technological dependenceBanks should map which AI providers and shared suppliers, if they failed, would affect the operations of a Hungarian bank.ECB says rapid AI adoption creates systemic risk in financial marketsConcentrated AI use and decisions built on shared models should be recorded as risks.76% of treasury teams hit by fraud as deepfake attacks risePayment-confirmation services for corporate clients should be updated according to deepfake risk.Consumer Reports: 9 in 10 Americans have encountered an online scam as AI fuels fraudIndustry bodies are also urging joint action against AI-driven fraud.Reserve Bank of Australia warns of AI bubble bursting in Financial Stability ReviewA price-correction scenario should also be considered when assessing exposures built on the AI market.
September 2026
Bank of England warns AI and debt risks together threaten financial stabilityCredit and capital-market exposure linked to AI should be tracked as a separate item in risk management.Hungarian surveys: one in seven Hungarians would already entrust their money to AICustomers' use of AI has reached financial advice, so it should be factored into customer communication and fraud prevention.Fed's Waller: AI shopping changes what banks need to authenticateFraud-detection and authentication models must distinguish between a transaction initiated by the customer and one initiated by their agent.UK launches AI scam watchlist as half of adults struggle to spot AI scamsFraud-prevention communication and victim support should be tailored to the characteristics of AI scams.UK financial regulators don't want AI-specific rules; FCA sees a leading role for UK in financeExisting regulation applies to AI use, without a separate AI law.AI-linked borrowers face a tough sell in riskier corners of the US credit marketFor AI-linked credit exposures, credit ratings and collateral need to be examined more strictly.Visa open-sources part of its AI-based cyber defence system, Portfolio reportsGiven vulnerabilities uncovered by AI models, patch and vulnerability management for payment infrastructure should be reviewed.Vanguard says AI financial advice is lifting demand for human advisorsIn advisory services, the human relationship needs to be presented as a value alongside AI.Survey: many finance chiefs weigh agentic AI capability when choosing a bankIn winning corporate clients, agentic AI capability is becoming a condition of competitiveness.Sony Bank and Fujitsu apply generative AI at full scale to financial system developmentGenerative AI is already a live, full-scale application in bank system development.Shopify opens checkout to AI agentsCard acceptance needs to be prepared for handling purchases initiated by agents.ECB to launch AI-agent payments proof-of-concept for the digital euro in JanuaryAgent-based use cases for the digital euro may require new payment processes to be built.Australian regulator to formally review banks' use of AI and its customer impactDocumentation and customer-impact assessments for AI tools used in customer communication should be prepared.Fed's Bowman warns banks on AI cyber risks while citing potential benefitsCybersecurity controls and system resilience should also be tested against AI-assisted attacks.Moscow finance firm A7 used AI-forged invoices to obtain banned technology, report saysIn trade finance, the authenticity of invoices must also be checked against AI forgery.Visa study: over 90% of European banks already use AIAI use is widespread among European banks, and competitive position needs to be measured against that.Morgan Stanley's Budapest hub has become an important international technology and analytics centreIn Budapest's financial technology roles, AI and automation are transforming tasks.Fraudsters use AI-cloned CEO voice to steal about €95 million from Intesa Sanpaolo's FideuramFor large transfers, an executive instruction received by voice or message must be confirmed through a separate channel by calling back.Mastercard expands Agent Pay with new trust and intelligence servicesAgent authentication and transaction context should be built into issuer-side decision rules.Stablecoin payment layer for AI agents takes shape: Verona, Jeeves, CloudflareBanks should assess to what extent agent micropayments will run alongside traditional payment rails.Apollo economist warns AI agents could trigger bank runs by shifting depositsDeposit-stability and funding-cost models should also be run against automated, yield-seeking behaviour.AI agents reportedly stole 600,000 credit card records at $25 per targetWith the cost of stealing card data plummeting, transaction monitoring and card reissuance policies may need tightening.CSBS announces AI supervisory framework for state-chartered banks and nonbank institutionsState-chartered banks need to assess whether their AI use meets the new supervisory expectations.Sibos 2026: banks move from AI copilots to autonomous workflows but leave judgment to peopleFor agentic workflows, banks need to define where human approval is required.Bank of America adds AI capabilities to its treasury client servicesAn AI assistant for corporate clients may become part of the customer service model.Gránit Bank offers new AI-based banking experienceHungarian competitors will need to respond with their own AI-based customer experience improvements.LHV Bank's UK arm rolls out agentic AI for customer emails with Gradient LabsEmail customer service may be handled by agentic AI, with levels of human oversight to be defined.War risk and insurance gaps threaten US data centres being built in the Middle EastThe resilience of banking services that depend on AI data centres must be assessed.Gránit Bank survey: bank customers more open to AI financial helpCustomer openness supports rolling out AI assistants, but trust and error handling are preconditions.Voice-cloning scam at Intesa Sanpaolo leads to €95 million in transfersTransfer instructions based on executive calls must be confirmed through an independent call-back.AI attack reportedly yields data of more than 600,000 bank cardsWhere card data is affected, issuers must check for signs of misuse and replace cards where necessary.Microsoft shuts down EvilTokens, an AI-based phishing servicePayment-fraud patterns may also rely on AI-based analysis of compromised email accounts.Feedzai launches Farol, an agentic AI for fraud investigation that cuts investigation timesAutomating fraud-investigation queues may reduce analysts' manual workload, but responsibility for decisions and auditability need clarifying.Frontier AI is testing operational resilience in European bankingOperational dependencies and failure handling of AI systems must be built into resilience plans.BlackRock: AI agents could drive stablecoin demand in machine-to-machine paymentsReplacing card and ACH rails for machine payments is being raised as a possibility, so strategy should account for it.FCA Mills Review AI findings; Claude for financial advisers with Schwab integrationResponsibility for deploying AI assistants built into advisory workflows needs to be clarified.Major banks warn that AI shopping bots raise fraud and data-privacy risksCustomer communication and fraud prevention need a separate scenario for unauthorized payments initiated by an agent.K&H prepares to accept payments initiated with Mastercard Agent PayAgent-initiated payments are also appearing at Hungarian banks, and the process for identifying and handling them must be developed.Global banks work together on principles for agentic commerceThe emerging industry principles may shape internal rules for agentic payments.Singapore's MAS reviews deepfake scam defences as cross-bank AI shield is testedBecause of deepfake-based fraud, authentication and confirmation processes need to be reviewed.AI could bring cost cuts and profit pressure to Hungary's financial sectorBanks should expect AI adoption to affect costs and headcount, while addressing regulatory concerns.FintechOS raises $28M after reaching profitability; Creatio to invest $300M in agentic AIThe range of vendors is widening, but buying agent-based solutions raises governance questions.Danske Bank and Mastercard complete Denmark's first agentic AI payment; Rogers Bank transaction in CanadaAgent-initiated transactions are already running live, so issuer-side handling and authentication should be tested.Report: AI chatbots mostly give wrong financial adviceBanks' own chatbots should be tested for the accuracy of their financial answers.New payment rails for AI agents: Polygon, Coinbase x402, Ripple MPP, Mesta, KredeteThe spread of payment rails outside the banking system challenges the traditional role of banks in payments.Visa, Mastercard and Ant build shared Know-Your-Agent framework as Sumsub links KYC to agent transactionsIdentifying the agent and proving human approval add a new authentication layer to the payment flow.Customer-service chatbots: UK told to keep human access; wrong-advice and insulting-reply casesAccess to a human agent must always be provided alongside the AI chatbot.JPMorgan's Dimon: AI infrastructure spending could approach $1 trillion in 2027The financing volume of AI infrastructure raises questions of credit and risk exposure.Danske Bank expects further job cuts as AI use expandsAI-driven efficiency gains are having a concrete headcount impact at a large Nordic bank.Oracle's Project Jupiter loans fall below 90 cents as SoftBank launches $10 billion-plus bond sale for OpenAI stakeTaking part in AI debt deals, or missing out on them, affects trading and investment banking revenue.AI-driven fraud wave: insurer complaints teams overloaded, Irish bank liability debate, deepfake investment scams in KazakhstanThe question of bank liability to victims of deepfake fraud is a legal and customer-handling risk.Fed's top bank supervisor focuses on banks protecting themselves from AI risksUS supervisors are paying increased attention to how banks manage AI risks internally.AI supervision frameworks: US state regulators, CFTC forum, Bank of England deputy governor's warningHuman oversight of autonomous AI systems needs to be built into documentable processes.S&P: AI use will soon influence banks' credit ratingsAI strategy and risk management will become factors in credit rating assessments.OCC conditionally approves bank charters for AI-agent and stablecoin infrastructureNew licensed competitors may enter agent payments and stablecoin services.Half of prop trading firms let staff use AI freely while banks set limits; AI-run funds emergeTighter AI limits at banks may mean a competitive disadvantage or lower risk compared with trading firms.Temenos and Celent: banks face a fight to keep three in four customers as expectations rise in the AI eraCustomer retention depends on improving AI-based customer experience and on how a bank's reputation appears through AI.Hungarian industry forums: who decides, who is accountable, who stops live bank AIDecision, accountability and shutdown powers must be clarified before going live.Content-provenance verification could help against deepfakes and fake newsVerifying the origin of content could add protection to customer and executive identification.Brazil tightens Pix rules over deepfake fraud; South Korea overhauls credit information consentDeepfake-related rules for instant payment systems could serve as a model elsewhere.IIF-EY annual survey covers AI use in financial services; PwC says AI skills are in higher demand than MBAsDeveloping AI skills moves up in career and hiring plans.JPMorgan rolls out Claude changes: $2,000 spending limits and extra securityA fixed limit and extra authentication on agent spending is a practical pattern for designing controls.US state bank examiners get a playbook for inspecting AIBanks under state supervision should keep an AI-use inventory and risk-management documentation ready.Payments modernization: 70% plan to deploy AI-based fraud detection within the next yearPayment system modernization is best scheduled together with the rollout of AI fraud detection.Two Michigan homebuyers lose $66,000 in suspected voice-cloning scamFor large property-related transfers, a separate call-back verification is warranted.Mastercard gives AI agents virtual cards, partners with AlchemyCard-issuing rules need to be worked out for transactions initiated by an agent.Lagarde: Europe faces unprecedented risk of being cut off from US AIDependence on critical AI providers should be treated as a supplier risk.AI-enabled fraud is 4.5 times as profitable as traditional fraud, reports sayFraud detection models must also be prepared for high-volume, AI-driven attacks.Visa, Mastercard and Ant seek to align AI agent verification, but disagree on the solutionThe standard for agent identification will determine what checks must be built into the payment flow.A quarter of financial institutions report deepfake or AI impersonation incidentsCustomer authentication must be supplemented with deepfake-resistant methods.Authorities and media warn of AI voice-cloning and investment scamsFor unusual transfers by elderly customers, consider applying a delayed approval.JPMorgan cuts lending to AI-focused hedge fund; hedge fund data quality and AI agent staff in focusPrime brokerage and lending limits for funds with AI strategies should be reassessed.Ant International launches protocol for AI agent paymentsA separate process is needed for handling agent payment instructions.Survey finds Hungarians increasingly open to using AI for their financesGiven customer openness, an AI spending-analysis or assistant feature is worth planning, with clear explanations, as customers are poorly informed.Sony Bank and Fujitsu apply generative AI to core banking system developmentGenerative AI is already a practical tool in core banking development, so a pilot is worth launching.JPMorgan private bank and Citi build AI into client relationships, but cautiouslyClient-facing AI should be piloted first, and the points of handover to a human must be defined.Dimon: AI oversight should be light touch; banks fix their own vulnerabilitiesVulnerability management and patching must speed up because AI-based flaw detection finds problems faster.AI agents have arrived in Hungary; one in ten online shoppers could use them by 2030A Hungarian bank must prepare its identification and approval processes for agent payments.MAS tests cross-bank AI models to detect scams fasterIn bank fraud detection, modelling data sharing between banks is a practical direction.Agentic AI in banking: six weeks after the EU AI Act, a US model risk update, SR 26-2The model risk classification of agentic AI should be aligned with regulatory expectations.Synapse Analytics raises $13 million for AI-powered decisioning at banksA new vendor option has appeared for banks' decision-making processes.Raiffeisen launches AI business coach in Romania; Libra Bank chatbot insults a customerA chatbot that talks to customers needs output filtering and an escalation rule.Chift raises €10.5 million; TCS to lead Aareal Bank's AI-powered transformationLarge bank transformation projects rely on external tech providers.SEON expands AI-based fraud prevention in the Philippines and adds signal intelligence against fraud ringsIn fraud detection, link (ring) analysis should be moved ahead of registration checks.JPMorgan curbs lending to Situational Awareness after AI losses, source saysExposure limits on financing and collateral management for AI-themed investment funds should be reviewed.OpenAI launches ChatGPT for Financial Services, aimed at Wall Street junior bankersThey need to assess which junior analyst and preparatory tasks the new product can replace, and where human review is needed.Survey: a quarter of Hungarians discuss their own finances with AICustomers often arrive after an AI conversation, so advice has to respond to the opinion they have already formed.Singapore to pilot common prefix for government calls and cross-bank AI against scamsCross-bank data sharing allows faster identification of suspected scam accounts, and the model can be adopted elsewhere.RBI warns of speed, concentration and opacity risks in AI use in financeFast, automated decisions need built-in human review points and a shutdown mechanism.Hungarian trade press: AI has opened a new era in fraud and makes it cheaperInternal payment-authorisation processes should require callback verification across multiple channels.Experian launches AI-enabled real-time decisioning platform in its consumer marketplaceCash-flow data can be a supplementary input for underwriting models.Chime to acquire Stride Bank, ending its sponsor-bank modelThe reshuffling of sponsor-bank relationships also affects partner banks' business.Ant International, Visa and Mastercard agree on common identity standard for AI agentsBanks need to prepare to identify, authorise and handle disputes for transactions initiated by agents.Axis Bank turns to AI to slow hiring and grow with a flat headcountEntry-level roles are shrinking, so training should focus on supervisory and review work alongside AI.Indian finance apps add AI agents: Super.money and L&T Finance with Pine LabsAgent-based payments in NBFC and app environments also affect transaction risk management.Gránit Bank and BME sign cooperation agreement, with AI ideas competitionHungarian banks are building AI talent and ideas through cooperation with universities.Canada's big banks use AI to recover loans and cut call-centre volumes as workers face uncertaintyCollections and customer-service roles are changing, and skills development should be shaped accordingly.Paytm to enter AI business, plans to sell agents to banks and insurersOn the vendor side, a new procurement alternative is emerging for voice-based and process-automation agents.Mastercard report: one in ten online shoppers could buy and pay through AI agents by 2030Banks need to prepare for agent-initiated card transactions and for handling customer authorisation.Mastercard launches Agent Connect for merchants; Visa Agent Pay is the other sideIssuers need to prepare to handle the two networks' agent tokens and identification.Czech and Slovak reports: scammers target people, with AI supportCustomer education and transaction warnings should be tuned to psychological manipulation and mobile-app attacks.KIWI Finance launches closed-environment agentic AI ecosystem for lending, built on FlowX.AIThe closed-environment agent system is a model for automating the lending process and managing its risk.True Balance scores every loan application with AI, using alternative dataIncluding alternative data can sharpen the assessment of customers without credit bureau profiles.Organised fraud rings use deepfakes and reused identities to target banksAccount-opening and customer-identification screening must be strengthened against patterns of synthetic and reused identities.India prepares AI-agent registry and Unified Agent Protocol for UPIBanks will need to verify agent authentication and handle authorised transactions.AI is live in about 170 places at Hungarian banks; payback and the human factor are the questionsAI pilots need to become dedicated projects with measurable payback, starting from staff use.FCA frontier AI review: financial firms inundated with AI-generated security findingsIT security teams must plan capacity to handle vulnerabilities uncovered with AI tools.Central bank makes AI a strategic priority; AI governance is the sector's taskFor AI in credit assessment, systems should be designed around the MNB's expectations and explainability.Anthropic unveils agentic commerce blueprint backed by Visa and MastercardBanks need to assess how purchases initiated through an AI assistant appear at the issuing bank.The AI return question: bank FOMO, CFO budgeting and a modest impact on resultsInvestment requests should be tied to a clear business metric.Revolut and Visa complete France's first AI agent-initiated paymentThe first experience of how agentic payments work on cards, and of customer approval, is worth analysing.Banks seek lead roles in Anthropic IPO and top credit ratings for Anthropic and OpenAIIPOs and credit ratings of AI labs open up a new client base for transactions and risk business.APRA and ASIC issue coordinated direction on frontier AI riskFor services built on frontier models, risk management documentation needs to be checked.FTC now targets payment processors that enable AI-driven fraudIn acquiring-side risk management, responsibility grows for screening out fraudulent merchants.Visa launches AI fraud shield meant to stop scams before money leaves accountsBanks should assess how pre-payment risk scoring could fit into their own fraud-screening systems.Claude, ChatGPT and Grok went down for an afternoon; none is listed among banks' critical providersA fallback procedure is needed for when the AI assistant goes down.Deepfake investment scams: Portuguese warning, Northern Irish and Indian victimsFor large investment transfers, a delaying verification question is worth applying.Fraud-prevention vendors: ITSS and Feedzai partner, Darwinium wins awardNew, award-winning players have appeared in the APP fraud solutions market and are worth comparing.Hungarian press: AI agents may haggle over prices and pay on users' behalfFor domestic payment products, the handling of agent-initiated transactions should be planned in advance.AI fraud pushes US toward cryptographic digital identityProcesses built on document-based identification should be reviewed.M&T Bank rolls out AI copilots to more than 15,000 employeesAlongside copilot use, a regulated scope of use and training should be planned.Incore Bank trials Gemini-powered KYC with KyndrylThe time saved through KYC automation is worth measuring in the customer onboarding process.Metropolitan Police: AI-related fraud reports up 395%, losses reach £9.6 millionFraud detection rules need to be tuned to AI-based impersonation and convincing messages.M&T Bank expands enterprise AI use after years of technology overhaulModernising the data and systems foundation is a precondition for AI adoption at banks.Bank Pekao tests AI that analyses customer feedback to improve its appAI analysis of customer feedback can provide a fast feedback channel for mobile banking product development.DC appeals court faults lawyers for Deutsche Bank subsidiary over AI hallucinationIn foreclosure and debt-collection matters, rules on AI use and verification need to be set with outside counsel.JPMorgan executive outlines what AI projects must prove to win internal fundingAI initiatives require a clear case for value and risk.Anthropic nears $15bn credit facility as Morgan Stanley and Goldman Sachs close to IPO rolesInvestment banks' roles in the listing and in lending could bring significant fee income.FCA warns frontier AI could create bottlenecks in fixing cyber vulnerabilities at UK financial firmsVulnerability patching priorities and capacity must be reviewed, because AI speeds up attacks.Santander and Mastercard report Europe's first payment made entirely by an AI agent; DBS and Stripe partnerAgent payments are already running in live banking pilots, so settlement and dispute-resolution processes should be prepared in advance.Visa upgrades its transfer-fraud prevention system with AI and Featurespace technologyWhen expanding transfer-fraud prevention, network-level, transaction-based scoring tools should be evaluated.Free family codeword app targets AI voice-cloning scams; Indian family business loses Rs 11.8 lakhA voice call alone cannot serve as reliable authentication for large transfers.Capitec's AI blocked $43 million in fraud losses in a yearThe return on fraud-prevention AI can be measured in losses prevented, which is useful for internal business cases.Mambu launches Intelligent Core to connect AI to core banking systemsThe AI integration capabilities of core banking vendors are becoming a new factor in decisions on replacing or extending systems.In Taiwan, households borrow and remortgage homes to invest in AI stocksHousehold lending carries growing exposure when the purpose of the loan is AI-stock speculation.Félix raises $200m to fund AI-powered product expansionLarge fintech funding rounds put pressure on banks' transfer pricing and speed.MNB recommendation and credit scoring: high-risk AI systems at Hungarian banksThe use of AI in credit-scoring models must be reviewed under the new MNB recommendation and the EU high-risk classification.India works on small-value UPI payments by AI agents under user-set rulesInstant payment systems need a new authorisation and limit-management process for agent-initiated payments.Revolut's AI assistant AIR launches in Hungary; data-handling fine print draws debateCustomer expectations of AI assistants at Hungarian banks are rising.Big banks and big tech clash over regulation of agentic AIThe outcome of the regulatory debate will determine who is liable for transactions initiated by an agent.Wealthtime: 28% of clients use AI tools to support financial adviceIn advisory conversations, expect that the client may already have received preliminary advice from AI.Banks lack the data to price in credit risk from AI-driven job lossesIn retail and mortgage lending, employment risks in AI-exposed sectors are hard to model.Australian AI firms get new partnership program backed by the big four banksThe program gives access to AI vendors for banks' development needs.Yettel customers can now ask their bank's AI about their spendingGrowing customer openness justifies faster rollout of AI-based customer service features.
August 2026
FSB chair warns G20 that frontier AI cyber risk threatens financial stabilityCyber-risk plans and vulnerability management must account for frontier AI models in attackers' hands.Visa, Mastercard and Fiserv join Agentic Payments Alliance as EMVCo publishes draft agentic payments frameworkComments on the EMVCo draft can be submitted until 30 September, and issuers should already consider their customer and agent identification logic.ECB tightens AI requirements for banks, giving them until October to adaptEuro-area banks must assess by October whether their use of AI meets the ECB's expectations.Anthropic unveils commerce agent blueprint with Visa and Mastercard; Mastercard launches agentic cohortNew network tools for protecting agent transactions should be included in the review of fraud-prevention architecture.Dutch banking association urges lenders to use AI for transaction monitoring; AML human oversight examinedFor AI-based monitoring, the rationale for alerts and the human decision points need to be documented.Deepfake video calls from the "bank" and a new face-verification bypass scam in HungaryFace-based identification alone is not enough; the process needs additional verification steps.MAS and industry publish SAFR framework for runtime safeguards on AI agents in financeThe SAFR control set can be a starting point for building runtime protection for agent systems.Germany: BNetzA to lead AI oversight, BaFin to address financial-sector AI risksAt German subsidiaries, it must be clarified which authority examines which use of AI.Goldman Sachs partner warns of 'huge danger' if AI replaces bankers' reasoning skillsWhen junior work is handed to AI, firms must ensure that analysts' judgment continues to develop.AI infrastructure on credit: ByteDance secures $29.6bn loan, Nscale closes $3bnLarge-scale lending for AI infrastructure raises questions of concentration and exposure.Hong Kong selects 36 AI use cases for financial-sector GenAI trials, including Ant BankThe selected cases show which GenAI applications supervisors are examining, including liquidity management.WhatsApp to screen suspected scam messages with on-device AINew protections on messaging platforms may change where scam attempts can be detected.Altruist launches AI financial planning agentAs AI agents are built into advisory workflows, the checking of recommendations and the information given to clients need to be regulated.German banking association comments on draft EU Cloud and AI Development ActCADA may affect the terms for procuring cloud and AI infrastructure, so the banking sector's position is worth following.Unico blocks R$19.6 billion in fraud attempts in first half, records wave of 'homemade deepfakes'Identity verification must allow for deepfakes now being made with home tools.Nvidia signs letter of intent with financial giants on $500 billion AI allianceThe role of large alternative asset managers in AI financing is competition for bank lending.Korean regulator to review banks' call centers over AI chatbot delaysAn AI chatbot must not come at the expense of access to human customer service.nCino: banks' prepaid AI allowance has run out, and they are buying moreThe cost of AI use and consumption of the allowance need to be measured.Entry-level adjuster hiring falls as insurers turn to AIAutomation of entry-level positions may also affect the skills people bring when entering the profession.Philippines' AMLC deploys AI to detect suspicious financial transactionsThe authority's analysis of reported transactions may become more advanced.Hungary's MNB reviews banks' AI sub-strategies as supervisors ask for evidence of AI operationsDocumented consistency must be established between the AI sub-strategy and actual operations.JPMorgan leads $5 billion debt financing for Volta's AI buildout as investor tolerance is testedFinancing AI infrastructure involves large-ticket credit risk that needs to be monitored.63% of advisers have clients using AI tools, and quality concerns persistIn client conversations, expect clients to bring AI-generated suggestions, and address them openly.DBS and Stripe partner on agentic commerce while Visa and Dunamu plan stablecoin AI paymentsAgentic and cross-border payment offerings should be designed together with partners.AI investment scams: New York issues official alert, losses reach $8 billionFor unusual transfers to a platform, consider warning the customer.Growhill Wealth launches AI automation tools for independent advisorsClient-management and documentation tasks can be automated, but professional responsibility remains.Mortgage lenders weigh AI, alternative data and credit modelsWhen using alternative data, model risk and the obligation to explain decisions must be managed.Scalable Capital opens platform to AI assistants; Fortune reports on AI trading at a German bankControl over AI assistants' access and the client's consent must be documented.Google Cloud launches Gemini Enterprise for Financial Services, with Deutsche Bank's inputManual research tasks can be automated, but checking the results remains a professional responsibility.Revolut sets up AI research unit, Revolut Research, for banking and fraud preventionDeveloping AI models against fraud requires in-house capacity.ECB tells eurozone banks to tighten cyber-security as AI changes the threat pictureCyber-risk controls and defences against AI-based attacks need to be reviewed.BaFin begins implementing EU AI rules in finance; ESAs issue statement on frontier AIAs the buyer of a credit-scoring or pricing system, the bank must ensure instructions for use and human oversight are in place.Goldman Sachs partner warns AI could erode bankers' independent thinking and the talent pipelineAI output must be something bankers can argue for and verify, or professional judgment will deteriorate.AI agents run 23 million stablecoin transfers in 30 days on x402 protocol as Payouts.com and Casper launch itMachine-to-machine payments are a new transaction channel, and products need to be prepared for it.Voice-cloning scams: banks respond as campaign says 28% of UK adults were targetedVoice authentication alone is not sufficient; additional checks are needed.Visa, Mastercard and 24 other firms form alliance to set standards for AI agent paymentsAuthorisation and dispute resolution for agent transactions should be designed in line with the card networks' direction.AI raises fraud risk at Australian banks as lending control gaps persistVerify income statements and identity documents against an independent source as well.Hungarian warning: a family code word can protect against AI voice-cloning phone scamsCustomer communication should include practical advice on fraud protection.Four major banks deploy Ant International's AI model for currency hedgingA pilot of AI models in treasury is worth considering.Bank of Japan: over 90% of Japanese financial institutions use generative AIAs an industry benchmark, generative AI use has become commonplace.JPMorgan files layoff notice in New Jersey as DBS reskills 11,000 staff for AIReskilling programmes and AI skills are increasingly a factor in keeping a job.Anthropic IPO could be led by major US banks at a valuation of up to $2 trillionThe large stakes of investment banking roles shape the capital-market position of the AI sector.TD Bank realizes $141 million in AI value, DBS scales agentic AI in corporate creditAutomating lending and corporate workflows can deliver measurable benefits.Kita raises $4.5 million to expand AI credit assessment for lendersLenders can also buy AI-based scoring from an external provider.Hungarian banks say AI will rewrite the world of work, leaving the value of a degree in questionIn banking careers, skills beyond a degree, including the ability to work with AI, are gaining value, so further training is worth planning.Starling launches AI-powered smart tools for personalised bankingCustomer-service and personalised AI tools give digital banks a UK competitor reference point.Citi, HSBC and Standard Chartered adopt Ant International's forex AI toolLarge banks are also using a third-party AI model for currency risk forecasting, so their model validation and risk management should be reviewed.JPMorgan adopts NJIT intern's AI assistant for email compliance; UBS hires AI headLarge banks are establishing AI leadership roles as distinct positions.92% of European banks use AI, but there is little public evidence of returnsFor AI projects, a measurement framework and baseline indicators must be set before deployment.Ten largest private banks cut more than 10,000 jobs in fiscal 2026 as AI reshapes hiringJob cuts are expected in tasks AI can replace, so a retraining plan is important.PPRO and BLIK bring agentic commerce to PolandAgentic use of instant systems like BLIK raises new fraud and limit-management questions.Voice-cloning scams spread: $25 million Hong Kong case, victims in India, advice on protectionFor extraordinary, urgent transfers, customers should be warned about the possibility of voice cloning.Asian banks must redefine what counts as an authorized payment in the age of AI agentsAuthentication and dispute-resolution rules must be amended for cases where an agent initiated the payment.Anthropic's pre-IPO credit facility could exceed $10bnLarge credit facilities for AI labs raise concentration and client-risk questions.Attacks start with $30 AI tools as the gap between fraudsters and banks widensBecause attack costs are falling, fraud-detection development must keep pace with attackers.Polish minister accuses Meta of 'darkest' censorship in deepfake scam disputeThe debate on platform liability affects how customer complaints over scam ads are handled.26 firms including Visa and Mastercard launch the Agentic Payments AllianceCard and account-management processes need to be prepared to handle agent-initiated payments.ECB researchers warn an AI-rally correction could hit the euro area hardLending and supply exposure tied to AI stocks needs to be reassessed.ASIC removes over 19,400 scam sites as deepfake investment fraud spreads in AustraliaWarning customers about suspicious investment transfers is a priority task in the era of deepfake ads.Coinbase launches AiFi after $100M in x402 transactionsBanks should expect part of agentic commerce to start on rails outside the banking system.Visa: issuer readiness goes live, but the agent economy lacks infrastructureCard issuers need to check that their systems recognise and authenticate trusted agents.Japan issues administrative guidance to tech giants over deepfake investment scamsScams built on deepfake ads show up as suspicious investment transfers, so screening for them is a priority.Morgan Stanley: European banks could cut 20% of jobs because of AIThose working in functions AI can replace should prepare for role changes and retraining.Indian banks may soon test AI defences against Mythos threats under RBI supervisionBanks should prepare to demonstrate, within a supervisory framework, their ability to defend against AI-based attacks.Hungarian government to deploy new tool against AI manipulation; 70% of corporate deepfake incidents are financial fraudThe government action may also affect fraud-prevention and customer-identification practice, and is worth following.BaFin takes over financial market surveillance under the EU AI ActClassifying lending and fraud-detection AI models as high-risk may bring supervisory review.Standard Bank becomes an AI-enabled organisation, with 39,000 staff active generative AI usersBroad staff usage is a benchmark: internal AI training and usage policy should be planned for similar coverage.GFF 2026: agentic AI and the reshaping of insurance underwriting and claimsBanks must plan for agents that carry out tasks, not only tools that answer questions.Study says finance needs AI-specific rules; new deepfake rules miss a key risk for banksBecause of gaps in the anti-deepfake rules, banks will have to rely on their own controls.Kenyan lenders turn to AI and alternative data for credit scoringScoring based on alternative data requires planning for model explainability and bias checks.African banks turn to AI against fraud as South Africa's sector faces mounting challengesThe spread of real-time risk assessment is pushing banks to reduce detection latency.South Korea moves to regulate AI risks in finance, with watchdog targeting hallucinationsGenerative tools used in customer communication will need a verification step and human review.Singapore: MAS SAFR runtime governance standard for agentic AI and a new AI taskforce for cyber resilienceLive oversight of agentic systems (logging, interruption) should be built in from the start.Money keeps vanishing from Hungarian bank accounts, and banks are powerless, Pénzcentrum reportsThe fraud trend reflects helplessness, so customer-warning and prevention processes should be reviewed.Visa adds AI to A2A Protect fraud prevention in EuropeBanks should compare their own fraud protection for account-to-account transfers with the Visa service at card level.Fraudsters abuse OTP's name with deepfake ads in 'OTP Casino' scamCustomer service must be ready for reports of such scams and provide proactive customer information.Colorado proposes automated decision-making rules; ABA calls for federal AI regulationReview automated decisions and customer chatbots against the new rules.FCA's Mills Review sets out vision for AI-enabled retail financial servicesTrack UK expectations when designing AI-based customer services.Hungarian analysis: MNB Recommendation 13/2025 and the AI Act shape AI practice in financeThe requirements of the MNB recommendation must also be enforced in supplier and partner contracts.AI and robots spread quietly through Hungarian banks, with AI complementing robotic automationStaff need to assess which of their processes can be automated and plan AI as an extension of robotic automation, not a replacement for it.AI chatbots are offering financial advice: how far can they be trusted?Customers increasingly bring chatbot-sourced advice with them, and advisers need to address it in their consultations.Gránit Bank survey: banking customers increasingly open to an AI assistantCustomer demand justifies building AI-based customer interfaces, with the option of human contact.Spanish police arrest man who used deepfake to bypass video identificationVideo identification should be strengthened with deepfake detection and additional checks.Visa cuts 2,600 jobs as AI reshapes its businessThe transformation of the payments network may affect partner relationships and service levels.RBI governor urges banks to speed up AI investment while keeping meaningful human oversightAI decisions need human review wherever an error could significantly harm customers or stability.Indian banks move from AI experimentation to deployment at scalePilot AI projects must be made scalable across lending, productivity and risk.AI-enabled fraud raises risks for banks in South Africa and the PhilippinesMulti-layered verification is needed to counter attempts to bypass deepfake-based identity checks.Autonomous AI agents breached a government network, with implications for banksPrepare for automated, AI-driven attacks and speed up vulnerability patching.European supervisors: banks must defend against cyber threats from frontier AIReview the IT security plan against AI-based attacks.OSL, Cloudflare, Orla and Sokin launch stablecoin payment products for AI agentsBanks need to assess how payments initiated by AI agents are to be handled on customer accounts.Deceptive phone calls on the rise, and fraudsters now target companies tooFraud warnings and customer education should cover corporate clients as well as retail customers.Moody's: AI race risks "systemic dependency" of big banks on tech firmsAssess the concentration of critical AI providers and the exit plans for them.AI infrastructure financing: shadow credit backstops, data center opposition and a new loan dealFor data center financing, environmental risk and community opposition are part of the credit assessment.DBS has around 26,000 AI agents running after a few monthsAI agents built by employees need a central register and access rules.Middle Eastern and Indian banks build on AI-native, agentic modelsAI-native banking models represent competition, so experience with deploying agentic services is worth following.Deepfake Ratan Tata investment scam in India; Guardian reports on AI scam exploiting holiday photosCustomer warnings should be extended to investment scams that use deepfake endorsements.AI fraud-detection tools from Yiren Digital, KFin and EDGE–Socure target lossesTesting fraud-detection and signature-verification AI tools is warranted to reduce losses.AI credit scoring: first-time borrowers in Egypt and legal gray areas across bordersFor AI credit scoring, data sources and cross-border legal requirements need to be clarified.JPMorgan Chase chief leads coalition on responding to increasingly powerful AIThe big-bank initiative may signal sector directions for managing AI risks.TSB research: more than half of those following social media financial advice lost moneyIt is worth providing young customers with credible, verifiable investment information.Moody's warns AI push is leaving banks dependent on tech firmsAI vendor concentration and exit options should be built into procurement and risk-management decisions.Gallup: few Americans trust financial advice from AIFor AI-based advice, building trust and signalling human review matter at the customer level too.Asian and Middle Eastern banks step up AI deployment: SBI, DBS, Ryt Bank, Vietnam, ruyaReplacing traditional processing with AI and AI-native bank models are appearing with increasingly concrete examples.ABN Amro partners with Mistral AI; Rabobank commits €2 billion to AILarge European banks are advancing with European model providers and big budgets, which is a benchmark in regional competition.JPMorgan's Jamie Dimon leads cross-industry alliance to tackle AI risksIt is worth watching the shared AI-safety practices that emerge from the alliance, as they may become an industry reference.AI-powered vishing wave hits Wall Street firmsPhone-based customer and internal identity checks should be supplemented with out-of-band verification.Gránit Bank survey: bank customers increasingly open to AI helpCustomer demand is worth factoring into development priorities for AI-based self-service and advisory features.Forecasts: AI agents could double financial fraudDetection rules and customer identification should be prepared in advance for agent-based fraud.Romania's BNR and ASF to audit bank algorithms under EU AI ActRegional subsidiaries should prepare to document their algorithms and make them auditable.PKO Bank Polski rolls out its first sovereign AI agentA large Central European bank has chosen a local, sovereign solution, which is a reference point for domestic data-management and vendor decisions.OTP Bank selects SambaNova systems to build Europe's fastest AI supercomputerA major Hungarian bank is investing in its own AI computing capacity, setting a pace and a supplier model in the regional competition.Cloudflare, Circle and OSL launch stablecoin wallets and payment layers for AI agentsStablecoin rails outside the banking system are competing with card- and account-based solutions for automated payments.Westpac puts AI agents into lending and opens AI lab; NAB to test agent safeguardsAgent-based lending and security testing of those agents is a working Australian example.Metro Bank customer fights for £14,000 refund after AI-linked fraudThe investigation and refund process after reports of unauthorised transactions is a supervisory and reputational risk.EBA, EIOPA and ESMA call for stronger governance of frontier AI ICT risksAI models and their providers must be brought into the ICT risk-management and third-party risk framework.Visa to buy fraud-defense platform BioCatch for $2.4 billionBanks that use behavioural fraud detection should assess how vendor dependence and contract terms may change under Visa ownership.Fed officials and analysts weigh financial-stability risks of debt-funded AI investmentLoan portfolios should be reviewed for exposure and concentration linked to AI-infrastructure financing.Visa, Mastercard and Corpay build trust rails and virtual cards for AI agentsAgentic transactions require separate identification and dispute-resolution processes.Financial firms cite AI in job cuts: Commonwealth Bank, Chime, BlockCustomer service and back-office roles are the most exposed; a retraining pathway is worth planning.New bank fraud spreads in Hungary: fake account noticesCustomer communications should inform clients about the current type of fake message.PSD3 and PSR: paying via AI agents raises new authentication questionsFor card and account payments, the handling of authority granted on behalf of an AI agent needs to be clarified.Plaid–Sierra and Sionic–Agentix link bank accounts to AI agentsThe consent and security model for account access by agents needs to be clarified.Deepfake fraud detection market report and warning for credit unionsDeepfake detection tools should be assessed for digital account opening and the lending process.EU AI Act delays give financial firms more time, not less responsibilityDocumentation and oversight of AI systems should be built up now.BaFin to supervise AI in the financial sectorInstitutions that also operate in the German market should expect AI supervision questions.South Korea to launch AI-based small-business credit scoring system in AugustState AI-based credit scoring may also affect the data sources and models used in lending decisions.