How Artificial Intelligence Is Helping Protect Customers From Fraud

Artificial intelligence is changing the way people work, communicate, shop, and manage their finances. It is also changing the cybersecurity landscape.

Fraudsters may use AI to create convincing phishing messages, imitate voices, generate fake documents, or personalize scams. At the same time, organizations can use AI and related technologies to identify unusual activity, analyze patterns, support fraud investigations, and respond to emerging threats.

This means AI has two very different sides. The same technology that can make fraudulent communications more convincing can also help strengthen the systems and processes used to protect customers.

Understanding this balance can help consumers and businesses use AI responsibly, recognize its limitations, and continue practicing the security habits that remain essential.

How AI Can Support Fraud Detection

Traditional fraud-detection systems often evaluate activity according to defined rules. For example, a system may flag a transaction because the amount, location, frequency, or other characteristic falls outside an established threshold.

More advanced technologies can analyze broader patterns and relationships across activity. The Federal Reserve Financial Services describes how financial-services fraud detection has evolved from static rules-based systems to predictive models and graph analytics. Predictive models use historical data to anticipate which transactions may be risky, while graph analytics can map relationships among people, accounts, and behaviors to identify patterns that individual transactions may not reveal on their own.

Generative AI is now becoming another part of that broader collection of tools. According to the Federal Reserve Financial Services, generative AI capabilities can work alongside traditional rules, predictive models, and graph analytics as part of a hybrid approach to protecting customers and organizations.  

In practice, AI and machine learning may help organizations:

• Analyze large amounts of activity more efficiently
• Detect patterns that warrant additional review
• Identify relationships across accounts, devices, or transactions
• Help investigators organize information associated with an alert
• Adapt detection practices as fraud tactics change
• Prioritize activity for further assessment

An alert does not necessarily mean that fraud has occurred. It may indicate that activity requires additional review or verification. Human oversight, established procedures, and appropriate security controls remain important parts of the process.

AI Can Help Identify Patterns People May Not Immediately See

A single activity may not appear unusual by itself. However, when several pieces of information are considered together, a broader pattern may emerge.

For example, the timing, location, device, transaction type, account history, or relationship among multiple activities may provide additional context. Analytical models can help process these signals at a scale that would be difficult to manage manually.

Graph analytics can also examine connections among people, accounts, and behaviors to reveal relationships that may not be evident when reviewing one transaction at a time.

This does not mean the technology is replacing human judgment. Instead, the technology may help direct attention toward activity that deserves review. Fraud and cybersecurity professionals can then assess the available information, follow established procedures, and determine the appropriate response.

AI Can Help Cybersecurity Teams Work More Efficiently

Cybersecurity teams often review large amounts of technical and operational information. AI-enabled tools may help organize that information, identify similarities among alerts, and support investigations.

These capabilities may allow specialists to spend more time assessing higher-risk situations rather than manually gathering and organizing every piece of information.

AI can also assist with recognizing patterns across phishing emails, fraudulent messages, account activity, and other signals. As scams become more personalized and scalable, efficiently reviewing information becomes increasingly important.

However, AI-generated findings should not automatically be treated as final decisions. Responsible use requires governance, oversight, testing, and an understanding of what the technology can and cannot determine.

The Same Technology Can Also Create New Risks

AI can support security efforts, but it can also help fraudsters create more realistic and scalable attacks.

Generative AI can be used to produce convincing phishing emails, scam messages, and social-engineering scripts quickly and in multiple languages. It may also be used to create synthetic identities, forged documents, or realistic deepfake audio, video, and images.

CNB’s own fraud-awareness materials recognize that AI can make social-engineering scams more realistic, adaptive, and scalable. Those materials identify AI-generated phishing messages, AI voice bots, and scripts that can change based on a person’s responses as emerging tactics.

These risks reinforce an important principle: a polished message, familiar voice, professional-looking document, or realistic video should not be accepted as proof that a request is legitimate.

Customers and businesses should continue to independently verify requests involving money, confidential information, account access, or authentication codes.

Responsible AI Use Begins With Protecting Information

Publicly available AI tools can be useful for brainstorming, research, writing, and other everyday activities. Before entering information into any AI platform, consider what information is being shared and whether the platform is appropriate for that use.

Avoid entering confidential information into a public AI tool unless the use has been approved and the platform is authorized for that type of information. This may include:

• Online banking credentials
• Passwords or authentication codes
• Account and payment information
• Customer or employee records
• Personally identifiable information
• Confidential business documents
• Internal financial information
• Proprietary business strategies

Consumers and businesses should also review the privacy and data-use terms of an AI service before sharing information. When in doubt, leave sensitive information out and use an approved, trusted resource instead.

Technology Works Best When People Remain Involved

AI can process information quickly, identify patterns, and support fraud-prevention work, but technology alone cannot replace sound security habits.

Customers and employees still play an important role by:

1. Reviewing account activity regularly
2. Protecting usernames, passwords, and authentication codes
3. Verifying unusual payment or information requests
4. Keeping devices and applications updated
5. Reporting suspicious communications promptly
6. Using trusted channels to contact a financial institution
7. Pausing before responding to an urgent or unexpected request

Businesses should continue using established payment approvals, vendor-verification procedures, access controls, and incident-response plans. AI can support these protections, but it should not be treated as a substitute for them.

A More Informed and Secure Digital Future

Artificial intelligence is creating both opportunities and challenges for cybersecurity. Fraudsters may use it to make scams faster and more convincing, while cybersecurity and fraud professionals can use advanced technologies to analyze information, recognize patterns, and support investigations.

The positive side of this change is that security tools can continue to evolve alongside the threats they are designed to address.

Customers do not need to understand every technical detail to benefit from these advances. The most important actions remain practical: protect confidential information, use strong authentication, verify unexpected requests, monitor account activity, and report concerns promptly.

By combining responsible technology with informed people and strong security practices, consumers, businesses, and financial institutions can work together toward a safer digital future.

Visit City National Bank’s Fraud Prevention & Security Center for additional resources and practical guidance to help protect yourself and your finances.

‍

Please note: The content in this article comes from individual opinions and experiences. The content should not be taken as advice coming from City National Bank of Florida. City National Bank of Florida does not offer tax, legal or accounting advice. CNBFL Member FDIC.