Fraud detection requires leveraging new tools and models to keep ahead of increasingly sophisticated fraud.
Financial institutions use AI to detect and prevent billions of dollars of fraud each year in areas including account takeover, check fraud, identity theft, money laundering, and terrorist financing, among others.
But they aren't the only ones. Increasingly sophisticated criminal organizations are also using AI to perpetrate financial crimes. And fraudsters are getting better and better at using it. In 2024, consumers lost more than $12.5 billion to fraud, a 25% increase from the prior year.[1] $3.1 trillion of illicit funds flowed through the global financial system last year.[2]
Monetary losses due to fraud can be significant, but the repercussions go beyond financial losses. Fraud can damage a financial institution's reputation. Customers become angry and frustrated when they're a victim of a financial crime or when a legitimate transaction is blocked. And they are not shy about sharing their experience on social media.
According to NVIDIA's 2025 State of AI in Financial Services report, financial institutions recognize fraud detection as a top use case (34%).[3] Accelerated data processing and advanced algorithms are improving AI's ability to detect and prevent fraud by identifying subtle patterns and anomalies in transactions, improving fraud detection accuracy and reducing false positives by 40%.[4]
"When a bank's balance sheet reserves anywhere from one to five percent for credit loss, there is tremendous opportunity to leverage AI to prevent fraud and loss," shared Malcolm DeMayo, Global Vice President - Financial Services Industry at NVIDIA. "That is why fraud detection is a priority use case for most firms."
How Fraud Detection Falls Short
AI uses multiple machine learning models to detect anomalies in customer behaviors and transaction patterns and identify potential fraud. However, these methods rely on rules-based systems or statistical techniques that are reactive and increasingly ineffective in identifying sophisticated fraudulent activities. Financial institutions need more advanced techniques to improve accuracy and reduce false positives, all in real-time.
In addition, the traditional computing infrastructure that financial institutions use often lacks the power to analyze the massive volume of transactions needed to identify abnormal behaviors and recognize patterns.
Criminals commit financial fraud in a variety of novel ways, such as hacking data from the dark web, using generative AI to create more human-like emails for phishing personal information, and laundering money between cryptocurrency, digital wallets and fiat currencies. FraudGPT writes malicious code and searches for vulnerabilities. To circumvent voice authentication, fraudsters use generative AI and LLMs to clone voices using deepfake technology.[5]
Beyond Machine Learning
Financial institutions' fraud detection solutions rely heavily on gradient-boosted decision trees, a type of machine learning algorithm. These algorithms are very good at identifying individual transactions or events that could indicate fraud. The problem is that criminals are increasingly hiding illegal activities within complex, connected networks.
However, AI is improving financial institutions' fraud detection efforts. Graph neural networks (GNNs) work with graph-structured data that analyzes the connections between accounts, transactions, and devices to reveal patterns of suspicious activity across the network. For example, even though an individual account doesn't raise red flags, GNNs can detect if that account is connected in some way to known fraudsters or unsavory entities.
Using AI-driven applications with GNNs, natural language processing (NLP), and computer vision, financial institutions can improve identity verification for know-your customer and anti-money laundering requirements, leading to improved regulatory compliance and reduced costs.
Generative AI can also produce synthetic data that speeds up fraud detection model training to keep ahead of the latest fraud schemes.
LLM-based assistants with retrieval-augmented generations (RAG) allow human workers to use natural language prompts to access vast datasets and speed up manual reviews.
Financial institutions can also use computer vision to analyze photo documentation such as drivers' licenses and passports to identify fakes. NLP can read documents to measure the veracity of the data on the documents and look for fraudulent records.
AI Factory Powers Fraud Detection
Fraud detection is complex and constantly evolving. Financial institutions struggle to keep ahead of emerging criminal schemes. However, AI models that incorporate tools such as GNN and RAG are a significant improvement over traditional AI and machine learning models. Enterprises need the computing power to handle large volumes of data and process it in real time.
"While we are still in the early stages of AI, it is important to make strategic investments to future proof your institution. Dell and NVIDIA are partnering so that a firm's investment today is protected down the road," added DeMayo.
The
About the Dell AI Factory with NVIDIA
Dell Technologies and NVIDIA can help you leverage AI to drive innovation and achieve your business goals. The Dell AI Factory with NVIDIA is the industry's first and only end-to-end enterprise AI solution*, designed to speed AI adoption by delivering integrated Dell and NVIDIA capabilities to accelerate your AI-powered use cases, integrate your data and workflows, and enable you to design your own AI journey for repeatable, scalable outcomes.
* Based on Dell analysis, July 2024. Dell offers solutions with NVIDIA hardware and software engineered to support AI workloads from PCs with AI-powered features and workstations to Servers for High-performance Computing, Data Storage, Cloud Native Software-Defined Infrastructure, Networking Switches, Data Protection, HCI and Services.
[1] https://www.ftc.gov/news-events/news/press-releases/2025/03/new-ftc-data-show-big-jump-reported-losses-fraud-125-billion-2024
[2] https://verafin.com/nasdaq-verafin-global-financial-crime-report/
[3] https://resources.nvidia.com/en-us-2025-fsi-survey/ai-financial-services
[4] https://www.forbes.com/sites/garydrenik/2023/10/11/generative-ai-is-democratizing-fraud-what-can-companies-and-their-consumers-do-to-prevent-being-scammed/
[5] https://www.ftc.gov/business-guidance/blog/2023/03/chatbots-deepfakes-voice-clones-ai-deception-sale