Fraud is where AI cuts both ways for a bank, and where regulators have been most specific. On the threat side, FinCEN, the FDIC, New York's DFS, the OCC and the ECB have all put in writing that generative AI now produces deepfake documents, voices and video that defeat identity checks and lowers the cost of attacks. On the defence side, the rules mostly encourage AI: the EU AI Act exempts fraud detection from its high-risk credit category, US agencies have encouraged AI in monitoring since 2018, and Treasury credits machine learning with $1 billion of its own fraud recoveries in a year. 93 of the 120 largest US banks profiled on this site record an AI fraud or scam-defence use case, 78 of them in production. What still binds: model validation in proportion to materiality, adverse-action duties when a fraud score declines a customer, and governance of the data the models learn from.
How banks use it
How are banks using AI in fraud detection?
AI in fraud detection at a bank means machine-learning models that score transactions, logins, applications and payments in real time, flag the anomalous ones and route them to a block, a step-up check or an investigator, and increasingly generative tools that help investigators work alerts. The bank records on this site show the range: machine-learning monitoring of payment flows at JPMorgan Chase and Wells Fargo, anomaly detection with behavioural and network analytics at U.S. Bancorp, generative AI for undefined threats at Truist, behavioural biometrics at Barclays US, agentic dispute handling at Old National and Seacoast, and customer education on deepfake and voice-clone scams almost everywhere. It is the AI use case regulators are most comfortable with: the EU AI Act carves it out of high-risk credit scoring and FinCEN's 2026 proposal counts effective use of AI in a bank's favour.
The largest with a recorded fraud use case: JPMorgan · BofA · Citi · Wells Fargo · U.S. Bank · Capital One · PNC · Truist. All 93 banks, with what each has disclosed ↓ Related: how fraud models are validated and AI chatbots in banking.
Bank examples
Fraud detection using AI in banking: which banks run what?
Eight bank records show the range. JPMorgan Chase describes machine-learning monitoring across roughly $10 trillion of daily payment flows (COO letter, April 2025) and Wells Fargo AI on payment flows of more than $1 trillion a day (2026); Bank of America counts more than 50 AI-enabled fraud-detection models (July 2025); Capital One's MACAW agents summarise long fraud calls for several hundred specialist agents (August 2026); Citi's Treasury and Trade Solutions screens payments with Feedzai; First Horizon writes AI detection rules for treasury payments in a day or two when a new scheme appears, and its CFO calls fraud the bank's number one AI use case (2026); FNBO uses Pindrop's agentic Fraud Assist, which Pindrop says cuts investigation time 35–40% (March 2026); American Express fraud teams are piloting AI that classifies incoming fraud claims (March 2026). Each row links to the bank's page, where the claim is sourced.
| Bank | Use case | What the record shows | Status | Latest source |
|---|---|---|---|---|
| JPMorgan | Fraud and AML analytics | Machine-learning monitoring across roughly $10 trillion of daily payment flows, per the bank's own operating description. | In production | Apr 9, 2025 |
| BofA | AI fraud-detection models | More than 50 AI-enabled fraud-detection models designed to detect fraudulent activity more quickly, accurately and at scale. | In production | Jul 16, 2025 |
| Wells Fargo | Payments fraud defence | AI on payment flows of more than $1 trillion a day; guidance to corporate clients on generative-AI fraud. | In production | Jun 1, 2026 |
| Capital One | MACAW fraud-call summaries | Understanding, reasoning, validation and explanation agents on fine-tuned open-weight models summarise long fraud calls for several hundred specialist agents; the same architecture runs Chat Concierge. | In production | Aug 13, 2026 |
| Citi | Feedzai payments fraud prevention (Treasury and Trade Solutions) | Feedzai quotes Manish Kohli, global head of payments and receivables at Citi Treasury and Trade Solutions, on Citi's 'strategic partnership with Feedzai', which he says helps it scale so clients can make payments 'securely, efficiently and without friction, across the globe'; Feedzai's 2021 funding release also named Citibank among users of its risk-management software. | In production | Oct 11, 2024 |
| First Horizon | AI-built detection rules for treasury-management payments | Rules are coded in a day or two when a new fraud scheme appears; matching outgoing payments are stopped and sent to the banker to call the client. The CFO calls fraud the bank's number one AI use case and says fraud is the one area still adding headcount. | In production | Jul 22, 2026 |
| FNBO | Pindrop Fraud Assist for card-fraud investigations | Agentic AI translates and summarises phone calls and generates case documentation for fraud analysts; Pindrop reports 35–40% less investigation time. | In production | Mar 17, 2026 |
| Amex | AI classification of fraud claims | Fraud teams are piloting AI technologies to analyse customers' fraud reports and classify incoming claims. | Pilot | Mar 25, 2026 |
Ranked by the same records: which large banks lead in AI fraud detection, and who leads in fraud and credit AI.
Deepfakes and voice clones
How can banks prevent deepfake and voice-clone fraud?
Deepfake fraud prevention starts from what the cases show. The AI incidents register on this site holds 22 cases, from 2022 to 2026, in which deepfake video, cloned voices or synthetic identities were used against a bank or its customers. In three controlled tests journalists' AI-cloned voices passed phone-banking voice ID: Lloyds in February 2023, Chase in April 2023, and Santander and Halifax in November 2024. The largest loss is at Fideuram, which wired about €95 million in February 2026 after fake messages from Intesa Sanpaolo's chief executive and a call in an AI-replicated lawyer's voice; at ABN AMRO a man opened 47 accounts in 2025 with deepfaked selfies and ID photos that the online identity check accepted. FinCEN's November 2024 alert sets out what supervisors expect: red flags for manipulated ID images and refused live checks, phishing-resistant multifactor authentication, live audio or video verification, re-verification of high-risk customers, and the SAR key term FIN-2024-DEEPFAKEFRAUD. The controls on the bank records are behavioural biometrics (BioCatch at Bank of America, Citi and Wells Fargo), phone-channel fraud scoring and passive authentication (Pindrop at FNBO), and device ID plus voice print on mobile wires (WaFd, October 2025).
| Date | Institution | What happened | Loss or amount |
|---|---|---|---|
| May 2026 | Lloyds Bank | Lloyds refunds estate of 82-year-old customer lured by an AI Farage deepfake investment scam | About £35,000 lost, including a £10,000 bank loan; fully refunded and the loan written off by Lloyds |
| Apr 2026 | Jio Payments Bank | Ahmedabad police: gang used Gemini-made deepfake videos to beat Aadhaar face checks and open bank accounts | Rs 25,000 loan in the first complaint; police estimated the racket earned Rs 10-15 lakh a year |
| Feb 2026 | Fideuram - Intesa Sanpaolo Private Banking and Intesa Sanpaolo | Fideuram wires €95m after fake Intesa CEO messages and an AI-cloned lawyer's voice | About €95 million transferred; about €53 million recovered (Il Sole 24 Ore: €59 million) and about €36 million missing |
| Jan 2026 | Klarna | Klarna processed payments for nine months for shops advertising with celebrity deepfakes | None reported |
| May 2025 | The Toronto-Dominion Bank (TD Bank Group) | TD customer, 88, loses C$450,000 to a scam that began with a Mark Carney deepfake; bank won't reimburse him | C$450,000 (CBC: more than $450,000 withdrawn from TD), virtually all of the customer's money |
| Mar 26, 2025 | HSBC (Singapore) | Deepfake Zoom call leads a Singapore finance director to wire US$499,000 from an HSBC account; funds recovered | Over US$499,000 transferred (about S$670,000); recovered |
| Mar 20, 2025 | ABN AMRO Bank N.V. | Man used deepfaked selfies and ID photos to open 47 ABN AMRO accounts in other people's names | €13,100 awarded to ABN AMRO for money it reimbursed to a bank-helpdesk fraud victim and for investigation costs |
| Feb 17, 2025 | Danske Bank | Criminals use AI-manipulated video and audio of Danske Bank's leadership to try to defraud the bank | None reported |
| Jan 21, 2025 | DNB Bank | Deepfaked DNB executives in a fake Teams meeting try to get Singapore staff to pay out millions | Close to NOK 24 million requested (several million Singapore dollars); nothing paid |
| Dec 7, 2024 | Kasikornbank | Thai court orders Kasikornbank to pay part of a loss to an AI face-swap fake-official video-call scam | 4 million baht lost (about US$118,000-124,000); bank ordered to pay 1 million baht plus interest |
| Nov 28, 2024 | Santander UK (Banco Santander) and Halifax (Lloyds Banking Group) | BBC reporter's AI-cloned voice passes Santander and Halifax phone-banking Voice ID | None reported |
| Nov 5, 2024 | PlainsCapital Bank (Hilltop Holdings) and Pathward, N.A. | PlainsCapital Bank customer sues after an alleged AI voice clone redirected a $20,000 transfer | More than $20,000 (about $21,008.68 claimed) |
| Feb 5, 2024 | OKX and OnlyFake | 404 Media passes OKX identity checks with a fake ID from 'neural network' site OnlyFake | None reported |
| Jan 2024 | Banxso (Pty) Ltd | Deepfake Musk and Rupert ads funnelled South Africans to Banxso; FSCA pulls licence and fines it R2bn | More than R115 million claimed by about 220 investors who contacted Moneyweb (October 2024); admitted liabilities above R137 million in the liquidation |
| 2024 | ING Groep N.V. | ING manager gets a fake voicemail from 'CEO Steven van Rijswijk' about a confidential project | None reported |
| Sep 2023 | bunq | Scammers deepfake bunq CEO Ali Niknam on a video call to ask staff for a large transfer | A "substantial" transfer was requested; the amount was not disclosed and nothing was paid |
| Aug 27, 2023 | Fortress Trust and Retool | Deepfaked voice in Retool breach linked to about $15m crypto theft at Fortress Trust | $12 million to $15 million in crypto (Fortress CEO to Fortune); CoinDesk put it at close to $15 million |
| Apr 28, 2023 | JPMorgan Chase (Chase) | WSJ columnist's AI voice clone passes Chase credit-card voice biometrics | None reported |
| Feb 22, 2023 | Lloyds Bank (Lloyds Banking Group) | Vice reporter breaks into his own Lloyds account with an AI-cloned voice | None reported |
| Jan 4, 2023 | Wells Fargo | DoNotPay says an AI voice clone got Wells Fargo to refund fees; the bank says the call didn't happen as shown | None reported |
| 2023 | Bank of America | AI clone of a client's voice calls a Bank of America banker to move money; attempt fails | None reported |
| Jul 2022 | Binance | Binance says scammers used an AI deepfake of its communications chief in listing-scam calls | None reported |
Related: banks with deepfake and voice-clone cases on record, voice biometrics and caller authentication at banks, FinCEN's deepfake alert and Pindrop.
Vendors
Which AI fraud detection software do banks use?
The AI tools directory on this site holds 23 fraud-detection vendors; 19 of them have at least one bank among the 120 largest US banks named publicly as a customer, each naming sourced on the vendor's page. The most-named are Nasdaq Verafin (8 banks), NICE Actimize (5 banks), Socure (4 banks). They split by where they score: transaction and payment platforms (Feedzai, Featurespace, NICE Actimize, FICO Falcon), digital-session behavioural biometrics (BioCatch), identity and onboarding (Socure, Alloy), the phone channel (Pindrop) and fraud-plus-AML platforms with AI investigation agents (Sardine, Hawk, Unit21). Most sell on sales-led enterprise contracts without published prices. The comparison with deployment and pricing is on the fraud-detection category page.
| Vendor | What it does | Banks named as customers | Ownership |
|---|---|---|---|
| Feedzai | RiskOps platform for bank fraud, scam and AML prevention: transaction scoring, behavioral biometrics and network intelligence. | Citi, SoFi | Private |
| Featurespace | Real-time behavioral fraud and AML scoring for banks and processors (The Featurespace Platform, formerly ARIC Risk Hub); a Visa unit. | U.S. Bank | Acquired |
| BioCatch | Behavioral-biometric and device intelligence that scores digital banking sessions for account takeover, scams, mules and application fraud. | Wells Fargo, BofA, Citi | Private |
| Sardine | Fraud and AML platform: device and behavioral signals, transaction monitoring, sanctions screening and AI agents for alert review. | None named publicly | Private |
| Socure | Identity verification, fraud scoring and watchlist screening for account opening and login, run on its RiskOS decisioning platform. | Capital One, Citi, SoFi, Atlantic Union | Private |
| Alloy | Identity and fraud prevention platform: onboarding KYC/KYB, fraud and login risk scoring, AML monitoring and data orchestration. | Ally, Banc of California, Eastern Bank | Private |
| Pindrop | Phone-channel fraud scoring, passive caller authentication and voice deepfake detection for bank contact centers. | FNBO, M&T | Private |
| LexisNexis Risk Solutions | Fraud, identity and financial-crime compliance data and analytics: ThreatMetrix, Emailage, IDVerse, Bridger Insight XG and Firco. | Cathay Bank, City National, EverBank | Business unit |
| DataVisor | Real-time fraud and AML decisioning platform using unsupervised machine learning, rules, device intelligence and Vera AI agents. | SoFi | Private |
| Oscilar | No-code risk decisioning platform with AI agents for fraud, credit underwriting, onboarding and AML compliance. | SoFi | Private |
| Prove Identity | Phone-based identity verification and authentication: Pre-Fill onboarding, passive authentication and AI-agent verification. | Synchrony, FNBO | Private |
| Mitek | Mobile check deposit capture, check fraud scoring and identity verification (document, selfie, liveness, deepfake detection) for banks. | JPMorgan, BofA, Capital One, U.S. Bank | Public |
| Jumio | Identity verification platform: ID and selfie checks, liveness and deepfake detection, AML screening and continuous identity monitoring. | None named publicly | Private |
| Persona | Identity verification platform: government ID, selfie and database checks, KYC/KYB, watchlist screening, case review and fraud-ring graphs. | None named publicly | Private |
| NICE Actimize | Fraud, AML, sanctions screening, KYC, case management and trade surveillance software for banks, on X-Sight and Xceed platforms. | KeyBank, Associated, Fifth Third, Valley, Cathay Bank | Business unit |
| Hawk | AML transaction monitoring, sanctions and customer screening, fraud prevention and AI investigation agents for banks and payment firms. | None named publicly | Private |
| Unit21 | Fraud and AML platform: transaction monitoring, screening, case management and SAR filing, with AI agents that run alert investigations. | Sallie Mae | Private |
| Nasdaq Verafin | Cloud financial-crime platform for banks: fraud detection, BSA/AML monitoring, sanctions screening, case management and AI agents. | FNBO, Pinnacle, Busey Bank, United Bank, United Community, OceanFirst, ConnectOne, First Interstate | Acquired |
| Zest AI | AI credit underwriting for banks and credit unions, plus fraud detection (Zest Protect) and gen-AI lending intelligence (LuLu). | Zions, First Hawaiian | Private |
| Provenir | Cloud decisioning platform for credit risk, fraud and identity, with a data marketplace, case management and AI models. | SoFi | Private |
| FICO | FICO Platform decisioning and analytics software for origination, account management and fraud (Falcon), plus the FICO Score. | Santander, TD, BMO | Public |
| Experian | Credit bureau and decisioning software vendor: Ascend analytics platform, PowerCurve decisioning, CrossCore fraud and identity checks. | Santander | Public |
| ACI Worldwide | Payments software for banks: real-time and card payment processing (ACI Connetic, BASE24) with AI fraud management. | BMO, First Hawaiian, City National, Webster | Public |
Deployment, pricing and the buyer's criteria: AI fraud-detection software for banks, compared.
What US regulators say
What do FinCEN, the OCC and the FDIC say about AI fraud detection in banking?
Each speaks to a different side of it. FinCEN supplies the typologies and the reporting duty: its November 2024 alert lists deepfake red flags and asks for the SAR key term FIN-2024-DEEPFAKEFRAUD, its September 2026 alert names 'artificial intelligence-supported software platforms' in the scam-centre ecosystem, and its April 2026 AML/CFT proposal counts 'effective use of artificial intelligence' in an institution's favour and says banks may adopt machine learning without added supervisory risk solely from adopting it. The OCC's Spring 2026 Semiannual Risk Perspective says AI is 'significantly transforming the cyber threat landscape', raising the speed, scale and sophistication of attacks and fraud, and recommends using AI defensively. The FDIC's 2024 Risk Review and its 2025 cybersecurity report to Congress say deepfakes and voice cloning make fake and synthetic identities harder to detect. The rules for the models themselves come from OCC Bulletin 2026-13 (SR 26-2) and the 2023 third-party guidance. The tracker holds no FFIEC document on AI fraud; its member agencies have spoken individually.
| Rule | Authority | What it requires | Applies |
|---|---|---|---|
| FinCEN Alert FIN-2024-Alert004 (deepfake media) | FinCEN | Red flags for manipulated or AI-generated ID images and refused live checks; phishing-resistant MFA, live audio or video verification, re-verification of high-risk customers; SAR key term FIN-2024-DEEPFAKEFRAUD. | Issued Nov 13, 2024 |
| FinCEN Alert FIN-2026-Alert005 (scam centers) | FinCEN | Names 'artificial intelligence-supported software platforms' in the ecosystem behind digital-asset investment scam centers; SAR key term FIN-2026-SCAMCENTERS. | Issued Sep 3, 2026 |
| FinCEN 2026 AML/CFT Program Proposed Rule | FinCEN | Effective use of artificial intelligence and federated learning counted among enforcement factors in a bank's favour; adopting machine learning or generative AI adds no supervisory risk by itself. | Proposed Apr 10, 2026 |
| OCC Semiannual Risk Perspective, Spring 2026 | OCC | AI lowers barriers for attackers and raises the speed, scale and sophistication of attacks and fraud; fraud and scams remain 'elevated'; recommends MFA, patching and defensive AI. | Published May 7, 2026 |
| OCC Bulletin 2026-13 (model risk management) | OCC | Non-generative fraud models are models: validation, ongoing monitoring and outcomes analysis scaled to materiality; generative and agentic AI out of scope. | In force from Apr 17, 2026 |
| OCC Bulletin 2023-17 (third-party relationships) | OCC | A vendor fraud model or platform is a third-party relationship: due diligence, contract terms, ongoing monitoring and exit, scaled to criticality. | In force from Jun 6, 2023 |
| FDIC 2024 Risk Review | FDIC | Generative AI used to get around identity and authentication controls with deepfakes, voice cloning and forged documents. | Published May 22, 2024 |
| FDIC 2025 Report on Cybersecurity and Resilience | FDIC | Deepfakes and voice clones make fraudulent or synthetic identities harder to detect at account opening, transaction processing and verification. | Posted Jul 14, 2025 |
What regulators are warning about
How is AI changing fraud against banks?
Regulators describe the same shift from three angles. Generative AI produces deepfake identity documents, voices and video good enough to defeat onboarding and authentication, which FinCEN, the FDIC and New York's DFS have all put in writing. AI lowers the cost and raises the speed of attacks, which the OCC, the ECB and DFS say changes the cyber threat landscape rather than any one scheme. And AI itself is used as bait, with trading bots and guaranteed-return schemes that the CFTC has warned customers about since January 2024.
| Rule | Authority | What it requires | Applies |
|---|---|---|---|
| FinCEN Alert FIN-2024-Alert004 (deepfake media) | FinCEN | Deepfake IDs, images, video and audio used to defeat identity verification; red flags, phishing-resistant MFA, live verification, SAR key term FIN-2024-DEEPFAKEFRAUD. | In force from Nov 13, 2024 |
| FinCEN Alert FIN-2026-Alert004 (federal student aid) | FinCEN | AI-generated synthetic identities and chatbots that keep 'ghost students' enrolled; SAR key term FIN-2026-FSAFRAUD. | In force from Jul 24, 2026 |
| FDIC 2024 Risk Review | FDIC | Generative AI used to circumvent identity- and authentication-based controls through deepfakes, voice cloning and forged documents. | Published May 22, 2024 |
| DFS AI cybersecurity letter | NY DFS | AI-enabled social engineering and AI-enhanced attacks must appear in Part 500 risk assessments, training and MFA design. | In force from Oct 16, 2024 |
| OCC Semiannual Risk Perspective, Spring 2026 | OCC | AI 'significantly transforming' the threat landscape: lower barriers to entry, more speed, scale and sophistication in attacks and fraud. | Published May 7, 2026 |
| ECB 'Dear CEO' letter SSM-2026-0301 | ECB | AI models that find vulnerabilities and generate exploits at unprecedented speed; action plans due to supervisors. | Plans due Oct 31, 2026 |
| CFTC AI scams customer advisory | CFTC | AI-branded trading bots, signal services and crypto schemes promising guaranteed returns. | In force from Jan 25, 2024 |
The rules AI fraud detection runs under
What rules govern a bank's own use of AI to detect fraud?
Fraud detection is the AI use case regulators are most comfortable with, and the rules mostly encourage it. The EU AI Act exempts systems used solely to detect financial fraud from its high-risk credit-scoring category; FinCEN's 2018 innovation statement and its 2026 AML programme proposal reward 'effective use of artificial intelligence' in monitoring; and US model-risk guidance treats a fraud model as a model, validated in proportion to its materiality. The constraints come from elsewhere: a fraud model that blocks or declines a customer can trigger adverse-action and UDAAP questions, and the data it learns from sits under BCBS 239 and privacy law.
| Rule | Authority | What it requires | Applies |
|---|---|---|---|
| EU AI Act, Annex III 5(b) carve-out | EU AI Act | AI used solely to detect financial fraud is excluded from the high-risk credit-scoring category. | In force |
| 2018 Joint Statement on BSA/AML innovation | FinCEN | Agencies encourage innovative approaches, including AI, to meet BSA/AML obligations; pilots will not be penalised for themselves. | In force |
| 2026 AML/CFT Program Proposed Rule | FinCEN | 'Effective use of artificial intelligence, federated learning, or other advanced monitoring tools' counted in favour of an institution. | Proposed; comments closed Jun 9, 2026 |
| SR 26-2 / OCC 2026-13 | Federal Reserve | Fraud models that are 'complex quantitative methods' are models: validation, monitoring and outcomes analysis scaled to materiality. | In force from Apr 17, 2026 |
| ECOA / Regulation B adverse action | CFPB | A fraud-driven decline of a credit application is still an adverse action needing specific reasons. | In force |
| ECB Supervision Newsletter on credit scoring and fraud | ECB | Decision trees dominate, neural networks used mainly for fraud, no self-learning after deployment; gaps on explainability and data standards. | Published Nov 20, 2025 |
| EBA report on AI adoption | EBA | AML/CFT and fraud detection among the dominant AI uses at EU banks; fraud alerts the most common consumer-facing generative-AI use. | Published Sep 25, 2025 |
What the public sector's own results show
Does AI fraud detection work at scale?
The best public number is Treasury's: machine learning that expedites identification of Treasury check fraud accounted for $1 billion of the $4 billion in fraud and improper payments prevented or recovered in fiscal 2024, across about 1.4 billion payments a year. The same Treasury found a 'fraud data divide' that leaves smaller banks without enough data to train comparable models, which is why data sharing and vendor models matter more to community banks than to the largest.
| Rule | Authority | What it requires | Applies |
|---|---|---|---|
| Treasury $4B fraud-prevention announcement | U.S. Treasury | $1 billion of FY2024 recoveries attributed to machine-learning check-fraud detection; $2.5 billion from prioritising high-risk transactions. | Published Oct 17, 2024 |
| Treasury AI cybersecurity risks report | U.S. Treasury | A 'fraud data divide' between large and small institutions; proposals for data sharing and vendor-AI 'nutrition labels'. | Published Mar 27, 2024 |
| Treasury AI in Financial Services report | U.S. Treasury | Sector-wide findings on AI use, including fraud detection, from the 2024 request for information. | Published Dec 19, 2024 |
By authority
Which fraud-related documents has each authority published?
Timeline
| Date | Event | Document |
|---|---|---|
| Jul 24, 2026 | FinCEN alert on AI-assisted student-aid fraud | FIN-2026-Alert004 (Federal Student Aid Fraud) |
| Jul 7, 2026 | ECB Dear-CEO letter on AI-enabled cyber threats | ECB 'Dear CEO' letter on AI-enabled cybersecurity threats (SSM-2026-0301) |
| May 7, 2026 | OCC: AI transforming the cyber and fraud threat landscape | OCC Semiannual Risk Perspective, Spring 2026 |
| Apr 10, 2026 | FinCEN proposal credits 'effective use of AI' in AML programmes | 2026 AML/CFT Program Proposed Rule |
| Nov 20, 2025 | ECB workshops on AI for credit scoring and fraud | Supervision Newsletter: AI use cases for credit scoring and fraud detection (Nov 2025) |
| Nov 13, 2024 | FinCEN deepfake media alert | FIN-2024-Alert004 (Deepfake Media) |
| Oct 17, 2024 | Treasury: $1 billion recovered by machine-learning check-fraud detection | Treasury $4B AI fraud-prevention announcement (Oct 2024) |
| Oct 16, 2024 | DFS maps AI-enabled social engineering to Part 500 | DFS AI Cybersecurity Industry Letter (Oct 2024) |
| Jul 12, 2024 | EU AI Act exempts fraud detection from the high-risk credit category | Regulation (EU) 2024/1689 |
| May 22, 2024 | FDIC Risk Review: deepfakes defeat authentication | FDIC 2024 Risk Review |
| Mar 27, 2024 | Treasury names the 'fraud data divide' | Treasury AI cybersecurity risks report (Mar 2024) |
| Jan 25, 2024 | CFTC warns customers about AI-branded trading scams | CFTC AI Scams Customer Advisory |
| Dec 3, 2018 | Agencies encourage AI in BSA/AML monitoring | 2018 Joint Statement on BSA/AML Innovation |
At the banks
Which of the 120 largest US banks use AI against fraud?
93 of the 120 bank pages on this site record an AI fraud, scam or payment-defence use case. Each links to the bank's page, where the claim is sourced.
| Bank | What the record shows | Status |
|---|---|---|
| JPMorgan | LLM payment validation screening: Large language models screen payment validations, reducing false positives and improving queue management; account validation rejection rates fell 15–20%. Federated-learning fraud detection (Kinexys): With BNY, RBC, DeepTempo and NVIDIA, Kinexys demonstrated federated learning that improves fraud detection while each institution keeps its data in its own environment. Fraud and AML analytics: Machine-learning monitoring across roughly $10 trillion of daily payment flows, per the bank's own operating description. | In production, Pilot |
| BofA | AI fraud-detection models: More than 50 AI-enabled fraud-detection models designed to detect fraudulent activity more quickly, accurately and at scale. Behavioural biometrics (BioCatch): American Banker (March 2026) names Bank of America among banks using BioCatch, whose technology reads how a user logs in, the device, the angle the phone is held at, keystrokes and past behaviour to decide whether the user is a real customer, a fraudster or a bot; Bank of America has not described its deployment. Person-to-person payment fraud analysis: AI has analysed person-to-person payments for fraud 7 million times, per the CTIO in November 2025; the bank says AI-enabled processes cut fraud losses in half since 2018. | In production |
| Citi | Behavioural biometrics (BioCatch): American Banker (March 2026) names Citi among banks using BioCatch, whose technology reads how a user logs in, the device, the angle the phone is held at, keystrokes and past behaviour to decide whether the user is a real customer, a fraudster or a bot; Citi has not described its deployment. Socure identity verification: Socure's company page lists Citi among organisations that use its identity-verification and fraud platform, and Forbes named Citibank among Socure customers in 2022; Citi has not described the deployment. Feedzai payments fraud prevention (Treasury and Trade Solutions): Feedzai quotes Manish Kohli, global head of payments and receivables at Citi Treasury and Trade Solutions, on Citi's 'strategic partnership with Feedzai', which he says helps it scale so clients can make payments 'securely, efficiently and without friction, across the globe'; Feedzai's 2021 funding release also named Citibank among users of its risk-management software. Cards controls monitoring: Cards uses AI as a 'smarter smoke detector' in controls, which Pam Habner said cuts the time to resolve customer issues by 95%. | In production |
| Wells Fargo | Payments fraud defence: AI on payment flows of more than $1 trillion a day; guidance to corporate clients on generative-AI fraud. Behavioural biometrics (BioCatch): BioCatch software that reads signals such as keystrokes, touch gestures and phone orientation to catch account takeovers, scams, money mules and fake account applications; Wells Fargo signed on in 2025, according to American Banker. Wells Fargo has not described the deployment. | In production |
| U.S. Bank | Treasury fraud defence: Anomaly detection, behavioural intelligence and network analytics on payments. Elavon Pay Defense: AI-driven merchant fraud and risk solution, named best risk, fraud and compliance solution at MPE 2026. | In production |
| Capital One | MACAW fraud-call summaries: Understanding, reasoning, validation and explanation agents on fine-tuned open-weight models summarise long fraud calls for several hundred specialist agents; the same architecture runs Chat Concierge. ProtectID dealer fraud screen: Proprietary fraud models and AI score every application sent through Dealer Navigator and flag synthetic and third-party fraud. Socure identity-fraud screening (U.S. Card): Socure says its identity verification raised approval rates and cut false positives for the Capital One U.S. Card team (vendor-reported, 2021); Socure's company page still lists Capital One among organisations using it. | In production |
| PNC | In-house fraud-blocking models: One of the named targets for the bank's own language models. | Announced |
| Truist | Fraud detection: ML for defined patterns plus generative AI for undefined threats. | In production |
| Schwab | Machine-learning fraud detection: Long-running ML models across the business. AI-powered fraud detection against AI-enabled scams: The eCrimes research and detection team is investing in AI systems that identify patterns and unusual behaviour at scale; a published patent application covers fraud detection during authentication. | In production |
| BNY | Federated-learning fraud detection: Proof of concept with Kinexys by J.P. Morgan, RBC, DeepTempo and NVIDIA: banks train local models and share encrypted updates instead of transaction data. | Pilot |
| TD | Fraud defence and claims: AI in fraud defence and insurance-claims fraud detection; $150 million claims-cost target. | In production |
| Fifth Third | Scam and impersonation defence: Expanded scam investigations; AI named by the fraud director as accelerating the threat. | In production |
| Huntington | AI-enabled fraud defence: Guidance and controls against AI-crafted BEC and impersonation. | In production |
| First Citizens | Fraud defence for business clients: Client guidance (a June 2025 webinar with McKinsey) on fraudsters' use of AI and on check fraud, recommending bank-provided controls such as ACH debit blocks and filters. AI in fraud and financial-crimes detection: The Chief Data and Analytics Officer lists fraud and risk management among the areas where the bank uses AI tools, and fraud and financial-crimes detection among the areas the in-house AI team led by Ash Kaduskar works on. | In production |
| Citizens | Payment-fraud defence: Head of fraud on AI deepfakes, BEC and bank impersonation. Agentic fraud-claims and AML checks: Fraud platforms rebuilt in the cloud with machine learning for suspicious-activity reports and security checks; agentic AI piloted for straight-through processing of small-dollar claims, with AML checks to follow. | In production, Pilot |
| Amex | Embedded fraud controls in B2B payments: Authentication and fraud controls in buyer–supplier flows. AI dispute-investigation pilot: Analyses historical transaction data so service reps can resolve disputes at once; AI and machine learning feed closed-loop fraud models. AI classification of fraud claims: Fraud teams are piloting AI technologies to analyse customers' fraud reports and classify incoming claims. | In production, Pilot |
| M&T | AI in fraud prevention and cyber defence: Jones says AI is used in cybersecurity, fraud and credit; Wisler names fraud prevention and cyber defence as early proprietary applications. | Rolling out |
| KeyBank | AI-deepfake fraud defence and client education: Guidance on AI-generated voice and video used in business email compromise. | In production |
| Northern Trust | ML detection of anomalous and fraudulent transactions: Listed in the 10-K alongside document digitisation and training services teams. AI in tokenisation security: Executives point to AI's role in detecting fraudulent behaviour around digital assets. | In production, Announced |
| HSBC | AI-redesigned fraud detection and credit applications: Among 50 processes under end-to-end redesign. | Rolling out |
| Regions | Biometrics and caller monitoring: A new biometrics tool and wider deployment of caller monitoring in the IVR to counter cyber criminals. Deepfake and AI-scam education: Client guidance on AI-generated impersonation. Adaptive authentication and identity intelligence: Authentication investment framed as reducing fraud and customer friction; the deck cites adaptive authentication, identity intelligence and continuous risk assessment. | In production |
| Pinnacle | AI-enabled fraud awareness for treasury clients: Guidance that fraudsters use AI too, with human review kept in payment workflows. Carefull elder-fraud monitoring (legacy Synovus): Vendor: Carefull. Its proprietary AI monitors bank, credit card and investment accounts around the clock for signs of fraud, scams and money mistakes; Synovus offered it as a complimentary benefit in a phased approach from February 2025. | In production, Rolling out |
| Santander | Deepfake-scam education and AI fraud models: Purpose-built fake AI ads in the UK; consumer education in the US. Synthetic fraud-network data (Santander AI Lab): gen-fraud-graph generates artificial networks of fraud transactions and behaviours, with no real customer data, for testing detection models. | In production, Announced |
| Western Alliance | Triangle of Fraud Protection (Digital Disbursements): Real-time screening and AI fraud scoring for class-action settlement payments. ClaimScore AI claim validation inside the Triangle: A third-party AI scoring engine supplies claim scores and real-time reporting to settlement administrators and feeds Digital Disbursements' screening; ClaimScore reviewed 60 million+ claims in 2024 and upgrades its fraud engine quarterly. | In production |
| Zions | AI in targeted fraud-detection workflows: Named by management as an early targeted use. AI for client authentication: Named alongside fraud detection and product recommendations as long-standing AI uses. | In production |
| Flagstar | Real-time core as the base for fraud decisions: CTO Pope says Finxact's real-time posting and single version of the data allow better fraud decisions and AI capabilities downstream; migrations are planned for summer 2027 and 2028. | Announced |
| First Horizon | AI-built detection rules for treasury-management payments: Rules are coded in a day or two when a new fraud scheme appears; matching outgoing payments are stopped and sent to the banker to call the client. The CFO calls fraud the bank's number one AI use case and says fraud is the one area still adding headcount. | In production |
| East West | Fraud-free access as the first call on technology spend: The CFO says investment has gone first to cyber, multifactor authentication and a fraud-free environment for client funds; no fraud-model or AI detail disclosed. | In production |
| Old National | Dispute and fraud-claim resolution (Quavo): Quavo, whose disputes platform includes agentic-AI features, gave Old National its client-satisfaction award; the bank has not disclosed which Quavo AI features it uses. | Announced |
| UMB | Client fraud-prevention education: Romance-scam and fraud guides for personal and business clients. | In production |
| SouthState | AI-fraud and synthetic-identity education: Customer guidance on AI-enabled scams. Voice and call fraud detection: Listed among call-centre AI uses; management cites AI productivity gains in fraud management. | In production, Rolling out |
| Columbia | In-house AI models on top of vendor fraud tools: Self-developed models layered on third-party fraud tools now catch fraud that would have bypassed them, the chief administrative officer said in September 2026. | In production |
| CIBC | AI in fraud detection and credit monitoring: Cited at the annual meeting as faster and more effective risk work. | In production |
| Valley | AI-powered fraud capabilities: Reached production with partners ahead of similar-sized banks. | In production |
| BOK Financial | AI-enabled fraud awareness and controls: Client guidance on generative-AI phishing, deepfakes and ACH/wire social engineering. AI in BOK's own fraud defences (CISO statement): Paul Tucker said in July 2025 that BOK is implementing AI in its systems to protect them while attackers use the same tools, and a treasury officer said in August 2026 that AI-powered tools help detect anomalies. No tool, vendor or result is named. | In production, Rolling out |
| Frost | AI fraud detection as the expected first use case: Laura Miller said fraud detection is likely to be one of the early AI use cases at Frost and most banks; no model, vendor or result has been disclosed. | Announced |
| FNB | Proprietary blend of 10 AI-enabled fraud tools: Shown at the November 2025 investor day alongside identity-theft controls such as liveness checks and device-integrity checks. | In production |
| SoFi | AI fraud investigations (Poetic): Vendor AI executes fraud investigations end-to-end with auto-decisioning; 99%+ quality reported within five weeks. AI/ML real-time payment-risk platform for clients: Technology Platform add-on that uses AI and machine learning for clients' payment-fraud mitigation, sold alongside a conversational AI engine. Socure identity verification and fraud prevention: Socure named SoFi a marquee customer in 2021 and a founding member of its First-Party Fraud Consortium in December 2024, and its company page still lists SoFi among organisations using it; SoFi has not described the deployment. Galileo Instant Verification Engine and Transaction Risk GScore: Real-time external-account and ownership verification plus a machine-learning risk score sold to Galileo clients; SoFi's risk platform now lists seven products whose models often use over 600 data points per decision. | In production |
| EverBank | Fraud prevention and risk tooling (Actimize, LexisNexis): In July 2024 the COO said Actimize fraud prevention software would launch by the fourth quarter of 2024 and that LexisNexis tools for fraud protection and risk decisions were already in place or about to be. The announcements do not describe machine-learning features and no later update is public. | Announced |
| Associated | AI-enabled fraud education and controls: Webinars and guides on social engineering, deepfakes and BEC. Deepfake and AI-enabled fraud defences: Harmening says pieces are already in place against video and audio deepfakes; client guidance centres on Positive Pay, Check Block and ACH Block, and Fiserv co-hosted a November 2025 webinar. | In production, Rolling out |
| Atlantic Union | Socure machine-learning identity verification: Listed in the bank's technology stack as an automated machine-learning identity-verification and fraud-management tool. AI fraud education for customers: A customer guide on generative-AI scams, including AI-driven phishing and cloned voices. | In production |
| Commerce | AI-scam awareness and investigations: Corporate investigations guidance on AI-enabled impersonation. Text alerts on suspicious checks: Fraud Detection Unit tools flag checks as they are processed and customers confirm by text; the bank does not describe the tools as AI. Client guidance on AI-driven payments fraud: Articles for commercial clients describe deepfake impersonation and generative-AI scams and point to AI anomaly detection and tokenization as defences; they are advice, not a disclosed internal system. | In production, Announced |
| Hancock Whitney | Payments-fraud policy position: FDIC response of 18 September 2025 asks for a federal fraud office, platform accountability and a Regulation CC fraud-hold exemption; no AI detection tooling is named. Client education on AI-enabled scams: March 2024 guidance covers AI-generated phishing, malicious chatbots, account takeover and AI-powered social engineering. | Announced, In production |
| BankUnited | AI-fraud awareness programme: Continuous client guidance on AI-enabled scams and AI-based fraud prevention. Commercial fraud-protection products: BankUnited's guidance names ACH alerts for transactions from unapproved parties, Payee Positive Pay for unmatched checks and Reverse Positive Pay, under which the client approves checks by email before payment; none is described as AI-based. The CEO reported a $10 million fraud loss on one commercial loan in the fourth quarter of 2025. | In production |
| FNBO | Pindrop Fraud Assist for card-fraud investigations: Agentic AI translates and summarises phone calls and generates case documentation for fraud analysts; Pindrop reports 35–40% less investigation time. AI-scam and deepfake awareness: CISO-authored guidance for consumers and businesses. Prove Pre-Fill identity verification for co-branded cards: Identity verification at application for the co-branded card portfolio, aimed at faster onboarding and lower fraud. | In production |
| Banc of California | Generative-AI fraud guidance: Client-facing guide to voice-AI, deepfake and LLM-written phishing and business-email-compromise scams. Generative-AI voice-authentication fraud exposure: Both the 2024 and 2025 10-Ks say generative AI makes fraud detection harder, particularly for voice recognition and authentication. | In production |
| United Bank | Customer fraud-awareness content: Fraudulent-website, phishing-email and domain-checking guidance. Card Controls alerts and lock in the mobile app: Customers get real-time alerts on unfamiliar or suspicious card use and can freeze or unfreeze a card. The bank does not say whether the alerts rely on machine learning. | In production |
| Texas Capital | Firm-wide agents in loan operations and fraud: In production by April 2026; 170+ processes mapped end to end (origination, approval, onboarding, monitoring) as candidates for further agents. | In production |
| Fulton | Customer education on AI-enabled scams: A bank article explains how AI-generated websites, reviews and videos are used in online retail fraud; it describes scams, not a Fulton detection model. | In production |
| Eastern Bank | Fraud and identity decisioning in the nCino lending platform (Alloy): Part of the consumer and mortgage expansion announced in November 2025; the 10-K separately warns that agentic AI makes synthetic-identity and deepfake fraud easier. | Announced |
| City National Florida | ACH and check Positive Pay, debit block and filtering: Rules-based screening of ACH and check items against issued payments, per-vendor debit filters, and SMS or email alerts; no machine-learning fraud engine is disclosed. Card Control in the mobile app: Cardholders switch cards off and set limits by location, spend, merchant and transaction type, with alerts for each transaction or attempt; no AI component is disclosed. Client education on AI voice-cloning and impersonation scams: A January 2024 post on AI voice cloning and a November 2025 alert on impersonation and account-takeover scams; education content, not a detection system. | In production |
| United Community | AI fraud-prevention tools: Credited with helping customers avoid potentially millions of dollars in losses. | In production |
| Arvest | Customer fraud education on AI-enabled scams: AI-powered phishing, deepfakes and synthetic identities; push-payment fraud; data quality. | In production |
| Ameris | Customer education on AI-generated fraud: Deepfake voices and AI-enabled phishing. Voice ID upgrade with automated risk checks: The upgraded phone system automates Voice ID risk checks and gives each call a risk score before it reaches a banker; the earlier Voice ID system protected more than $95 million of customer funds. The report does not call this AI or name a vendor. Photo ID scanning to flag fraudulent IDs: A scanning system added to strengthen protective measures at account opening; the vendor and method are not disclosed. | In production, Rolling out |
| WaFd Bank | Customer education on deepfake scams: Voice cloning and AI impersonation guidance. Voice-print verified consumer wires: Consumer wires in the mobile app are verified by device ID plus the client's voice print for $10; launched with Release 2.8 in October 2025. | In production |
| WesBanco | Client education on AI-generated fraud: Deepfake and account-takeover guidance for businesses. | In production |
| Renasant | AI-enabled fraud named as a threat: The 10-K says AI has made it easier to mimic customer communications and signatures and to create synthetic instructions, documents and media; controls are described as preventive and responsive. | In production |
| Customers Bank | Qualia Shield wire-fraud verification for title and escrow: Wires originate and are received inside Qualia, with automated fraud verification on every transaction; Qualia is described as an AI-powered closing platform. AI fraud and deepfake detection (stated): Customer guidance says the bank uses real-time fraud detection powered by AI, biometric authentication and deepfake and impersonation detection; no vendor, model or metric is disclosed. | In production |
| Rockland Trust | AI-assisted fraud review: Fraud review is one of three named back-office targets of the AI innovation team (22 January 2026 call). Digital Operations and the BSA/Fraud Financial Crimes Unit sit under Debra Smith. No model, vendor or result is disclosed. | Announced |
| Simmons Bank | Mobile-deposit customer scoring: In-house scoring and qualification algorithm built with Jack Henry: evaluates every customer daily, recalculates limits monthly, sets deposit eligibility, limits, risk level and holds; fraud losses down 96% to under 1 basis point; the source does not say whether the model is machine learning. | In production |
| First Hawaiian | Deepfake and scam awareness for customers: Deepfake protection at work and home; malware and job-scam warnings. AI voice-cloning scam education: How voice-cloning scams work and how families can protect themselves. Planned Zest Protect fraud detection: Zest says the bank is considering adding its AI fraud-detection product to the lending process and a GenAI lending companion (LuLu); the bank has not confirmed either. | In production, Announced |
| Cathay Bank | Positive pay and account alerts (no AI described): The 2025 annual report cites positive pay features for fraud protection on client trust accounts; the bank's client guidance covers customised account alerts and does not mention AI-based detection. | In production |
| Bank of Hawaii | Scam education for seniors: Top-ten scams guidance. AI-assisted scam awareness: Guidance on how AI tools help scammers exploit social-media posts. | In production |
| Centennial Bank | Loan-origination fraud framing in the 10-K: The 10-K names appraisal and underwriting-document misstatement by third parties as the largest fraud risk, and lists ACH, wire, ATM and checking transactions as exposures; no machine-learning detection is described. | In production |
| First Financial Bank | Commercial fraud education including AI deepfakes: Annual checklist and commercial fraud hub for business clients. AI fraud seminars for business clients: The bank's 'Fraud in the Age of AI' seminar combined a live deepfake demonstration with a panel of the CISO, a Clark Schaefer Hackett IT-risk shareholder and a Secret Service agent, moderated by the director of enterprise fraud. The commercial fraud hub lists 'AI and deepfakes' among four fraud trends. | In production |
| TowneBank | Customer education on AI scams: Voice cloning and deepfake awareness. | In production |
| Mechanics Bank | Customer fraud education: Impostor scams, tax-season fraud and senior fraud prevention. | In production |
| Seacoast Bank | Agentic AI dispute management (Quavo QFD): Intake to resolution automation with AI fraud detection trained on millions of cases. Proactive outreach on atypical account activity: A market president says the bank uses technology for early identification of fraud, often before the customer knows, and contacts customers whose activity departs from their normal trend. | In production |
| First Merchants | Customer education on text scams: Recognising smishing and impersonation. Positive Pay for business clients: Check Positive Pay matches account number, check number and amount against issued-check files; ACH Positive Pay adds pre-approved authorisations and transaction-monitoring rules; clients decide exceptions by 3:00 p.m. Eastern. Rules-based, with no AI claimed. | In production |
| Optum Bank | AI screening for high-risk and deep-fake callers: Optum Financial says it uses AI and audio identifiers to flag risky calls and deep-fake callers and to verify legitimate callers faster. No vendor or accuracy figure is disclosed. | In production |
| Central Bank | Customer education on AI-enhanced scams: Voice cloning and impersonation guidance. | In production |
| Merchants Bank of Indiana | Wire-fraud customer education: Guidance on verifying payment instructions. Mortgage-fraud exposure in multifamily lending: Mortgage fraud or suspected fraud by borrowers contributed to a $43.1 million rise in the Q2 2025 provision and $46.1 million of charge-offs; management said it was enhancing risk management practices. No detection technology is named, and the 10-K lists fraud detection only as an industry AI use case. | In production |
| Trustmark | Positive pay for business clients: Trustmark Bank said in May 2025 that it had implemented multiple types of positive pay for check and ACH fraud; the vendor release describes rule-based payee and amount controls, not machine learning. | In production |
| Bank of Hope | Positive Pay and ACH Transaction Review (business clients): Payee Positive Pay compares checks with the issue file, ACH Positive Pay blocks debits from companies the client has not authorised, and ACH Transaction Review confirms postings case by case; rules-based, and the bank does not describe them as AI. Customer-set card controls: Consumers block or limit cards by merchant category, transaction type (ATM, e-commerce) and amount, with real-time alerts, in online banking and the app; no AI is disclosed. | In production |
| Busey Bank | Customer education on AI fraud and deepfakes: Guidance for consumers and businesses. AI-powered fraud detection (Verafin named in the FY2024 10-K): The FY2024 10-K cites 'artificial intelligence-powered tools, such as Verafin' as additional layers of fraud detection; the FY2025 10-K says AI-powered tools provide layers of fraud and threat detection without naming a vendor, so the current vendor is not confirmed. | In production |
| Community Bank N.A. | Customer payment-fraud and cybersecurity risk software: The 2025 10-Qs attribute rising data-processing expense to software for customer payment fraud and cybersecurity risk management alongside AI applications; the filings do not say whether the fraud tools use machine learning. | In production |
| Enterprise Bank & Trust | Annual fraud-protection guidance: Payments and fraud controls for business clients. AI-fraud awareness for business clients: October 2025 guidance on voice and face cloning, AI-written phishing and fake IDs; recommends multi-factor authentication, known-number call-backs, code words and phishing simulations, and Treasury Management fraud tools. It describes criminals' use of AI, not a bank detection model. | In production |
| FirstBank | Commercial customer fraud controls: Guidance of 16 September 2026 points business customers to dual control, multi-factor authentication, Positive Pay with Payee Match and daily account monitoring; the page does not describe AI or machine-learning detection. | In production |
| First United Bank | Business fraud education: AI-generated fake invoices: Guidance to train staff, update cybersecurity systems and review vendor-payment protocols. Business Fraud Risk Calculator: Self-assessment tool for business customers built on the bank's account data and AFP statistics; companion guidance recommends AI-driven anomaly detection and staff training. | In production |
| Banner Bank | Fraud-related technology investment: CFO Rob Butterfield said in October 2025 that fraud costs outpace credit losses and that Banner keeps investing in fraud-related technology; the 10-K warns that AI disruptions could affect AI-supported fraud detection and transaction monitoring. No fraud vendor or model is named. | Rolling out |
| Synchrony | Behavioural fraud detection and AI-enabled fraud defence: A credit executive says the company uses data to learn what is typical for an individual applicant or customer, drawing on device identifiers, transaction patterns and external data; the 10-K warns fraudsters may use generative AI. Cashflow analysis to identify credit washing: Consented cashflow analysis in PRISM helps the bank identify fraud such as credit-washing schemes that borrowers use to improve their scores. | In production |
| USAA | AI to detect and prevent some types of fraud at the bank: USAA's AI page says the bank uses AI to automate processes, enhance member service and detect and prevent some types of fraud; AI-powered fraud detection is also cited for claims. Quavo QFD agentic fraud and dispute platform: USAA Savings Bank selected the platform in September 2025 to automate disputes from intake to resolution; Quavo says its Aria agent works through investigation steps and proposes resolutions. No go-live date or result is public. | In production, Announced |
| Popular | Customer education on AI-enabled fraud: Popular Bank's security pages warn customers about deepfake impersonation, robocalls and account takeovers; the page does not describe the bank's own fraud models. | In production |
| Wintrust | Enterprise fraud management and transaction monitoring: Wintrust runs a fraud transaction monitoring team under Tim Murphy and an enterprise fraud management function under Ray Olsen; Olsen's team monitors the dark web for stolen Wintrust checks and bankers contact customers before acting on unusual activity. The 10-K says Wintrust may selectively use AI in fraud detection; no filing or interview says which fraud models use AI. | In production |
| OceanFirst | AI for fraud and BSA review, AI-driven KYC/AML: Listed in the 2026–2027 scale-and-integration phase of the automation roadmap alongside personalised next-best products. | Announced |
| Citizens Business Bank | AI flagging of departures from known transaction patterns: The bank and certain vendors use AI tools to enhance the security of certain customer banking transactions by tracking and flagging departures from known patterns. Customer education on AI-enabled scams: The Security Center page is headed 'Fraud is Getting Smarter with AI' and links a customer cybersecurity newsletter series whose Volume 2 covers AI-driven scams and whose Volume 4 covers bot-driven fake accounts on social media. | In production |
| FirstSun | Fraud risk named in the 10-K: The bank warns that recent advances in AI may make fraud more difficult to detect, including generative-AI social engineering; it has not described AI-based fraud defences. | Announced |
| Nicolet | AI as a fraud accelerant: The 10-K warns AI and quantum computing could exacerbate card, check and social-engineering fraud; no AI-based defence is described. | Announced |
| ConnectOne | AI in fraud detection: The 10-K says the bank has begun to selectively incorporate AI in certain business processes and fraud detections; it does not name the models or vendor. The CEO separately names Verafin as a vendor adding AI to its platform. | Rolling out |
| Bread Financial | AI-enabled fraud and cyber threats: The 10-K says generative AI, including synthetic voice, adds identity-fraud risk and that threat actors use AI against security controls; fraud-related operational losses were $65 million in both 2025 and 2024. Proprietary fraud models including machine learning: The 10-K says Bread uses proprietary custom fraud models built by its data scientists plus externally sourced scores, device intelligence for digital applications and machine-learning models on monetary transactions, alongside rules, report analysis and manual review. | In production |
| BancFirst | Debit-card fraud detection programs: The 2024 sustainability report says BancFirst runs 'various fraud detection programs to search for and alert customers of potential fraud' on its debit cards. It names no vendor and does not describe the programs as AI. | In production |
| National Bank Holdings | AI-enabled fraud as a threat: The 10-K says the widespread use of AI has increased potential for fraud and misuse and that the bank maintains fraud detection tools and training; it reports limited fraud losses. Fraud controls in 2UniFi onboarding: John Finn said 2UniFi prioritises a high-quality onboarding experience with robust fraud mitigation; the 10-K separately says NBH maintains fraud detection tools and training and has had limited fraud losses. NBH does not say these controls use machine learning. | In production |
| First Commonwealth | AI-enhanced fraud as a threat: The 10-K says emerging technologies such as AI may further enhance criminals' ability to perpetrate fraud despite its monitoring and fraud-detection systems. Customer education on AI deepfake fraud: The bank's fraud-prevention pages warn consumers and businesses about AI-generated voice and video impersonation, synthetic identities and AI-driven phishing, and advise multi-factor authentication, call-backs and dual authorisation. The pages describe customer-side threats, not a bank AI detection system. | In production |
Common questions
Is AI fraud detection regulated differently from other bank AI?
It is treated more leniently. The EU AI Act excludes systems used solely to detect financial fraud from its high-risk credit-scoring category, US agencies have encouraged AI in BSA/AML monitoring since 2018 and FinCEN's 2026 proposal counts effective use of AI in a bank's favour. What still applies: model-risk validation in proportion to materiality, adverse-action duties when a fraud score declines a customer, and data governance over the training data.
What are regulators warning banks about?
Deepfake identity documents, voices and video that defeat onboarding and authentication (FinCEN, FDIC, DFS), AI that lowers the cost and raises the speed of cyberattacks (OCC, ECB, DFS), synthetic identities in benefit fraud (FinCEN), and AI-branded investment scams aimed at customers (CFTC).
How many of the largest US banks run AI against fraud?
The bank pages on this site record an AI fraud or scam-defence use case at half of the 100 largest US banks, most of them described as in production; the table below lists each with its source.
Where does the fraud data divide bite?
Treasury's 2024 report found smaller institutions lack the data to train their own anti-fraud models, which pushes them to vendor models and consortium data. Third-party risk guidance (SR 23-4) then applies to the vendor relationship, and the bank remains responsible for validating what it buys.
Fraud in the daily brief
Nov 2, 2026DOJ's Operation Ghostbusters charges five in $200 million ghost-student loan ringOct 30, 2026NCUA plans $4.3 million AI investment for fraud detection, supervisionOct 29, 2026Bank of America asks customers not to share passwords with AI assistantsOct 28, 2026Delaware plan would let AI agents run entire businessesOct 27, 2026First Internet Bank cuts SBA compliance reviews 87% with AI agentsOct 25, 2026SEC's 2027 exam priorities put RIA conflicts and cyber risks in focus
Navigate
Every new fraud alert and every bank's fraud move, the morning after.
when one of these regulators moves, the next morning's brief says so · six sourced stories · 7 am PT · free
plus every regulation, bank and agent page update, the morning after · leave any morning