Financial institutions are facing a significant surge (Off-site) in synthetic identity fraud as criminals increasingly leverage artificial intelligence (AI) (Off-site) to scale and streamline their operations.
Synthetic identity fraud is the use of a combination of personally identifiable information (PII) to fabricate a person or entity in order to commit a dishonest act for personal or financial gain. Check out the Synthetic Identity Fraud Mitigation Toolkit for insights and downloadable resources on addressing this type of fraud.
Furthermore, generative AI can create realistic text, images, video and audio — and when combined with malicious code, allows criminals to automate key steps in synthetic identity schemes, including:
- Creating synthetic identities and producing convincing identification and supporting documents (such as utility or mobile phone bills)
- Submitting applications for new bank accounts and credit cards
- Satisfying security measures, including facial recognition and liveness detection, during account opening
- Conducting transactions at scale across multiple institutions to mature accounts, strengthen profiles and increase available credit
These generative AI tools are readily accessible and often free or low-cost, giving bad actors the ability to rapidly expand their operations. This is not a distant or hypothetical threat, as it is already reshaping the fraud landscape. According to a recent report (Off-site), synthetic identity fraud continues to increase year over year and was the fastest-growing type of fraud globally in 2025, with an eight-fold increase from 2024. This level of automation enables criminals to conduct end-to-end customer interactions on a large scale, significantly increasing both the volume and sophistication of synthetic identity attacks. To help mitigate potential losses, financial institutions must understand how emerging AI tools enable synthetic identity fraud and assess the effectiveness of their fraud prevention controls for new accounts and credit products.
How Generative AI Creates Synthetic Identities
Synthetic identity fraud is the use of a combination of personally identifiable information (PII) to fabricate a person or entity in order to commit a dishonest act for personal or financial gain. AI systems can gather massive amounts of data — names, addresses, Social Security numbers, dates of birth, employment histories — from sources such as data breaches, public records and social media. Generative AI then uses this information to craft highly believable identities that mirror legitimate profiles. What makes these identities especially dangerous is their potential hybrid nature. Criminals often blend authentic data points, such as a real Social Security number or physical address, with fabricated details generated by AI tools. They also can use a combination of entirely real or fabricated information. Additionally, AI tools can create supporting documents, such as driver’s licenses, passports and utility bills, accurately reproducing government markings, security features and business logos. The result is a plausible yet nonexistent identity that can pass early-stage verification and screening. This mixture of real and synthetic elements makes these profiles extremely difficult for traditional systems to detect and enables criminals to open accounts and obtain credit.
Using Synthetic Identities to Open Accounts
AI-driven deception has challenged the account opening process. Deepfakes — AI-generated images, videos and selfies — now can satisfy identity verification requirements for credit cards, loans and bank accounts. These creations so convincingly mimic real voices, facial movements and behaviors that both automated systems and human reviewers can be fooled.
But visual deception is just one piece of the threat. Generative AI-powered bots can engage in natural-sounding conversations, navigate verification workflows and execute social engineering tactics at scale. By automating identity creation, documentation and communication, criminals simultaneously can submit multiple applications across numerous institutions. This level of automation drastically reduces the time, cost and expertise required to commit synthetic identity fraud and makes it significantly harder for financial institutions to identify fraudulent accounts during onboarding.
Test your Knowledge: Real or AI-Generated Images
Can you distinguish a real photograph from an AI-generated image? Review the images below and determine whether they are real or AI-generated. As you examine the images, consider what indicators might help you decide, such as unfinished forms, blurry backgrounds or near-perfect images. (See answer key at the end of this article.)
Were you able to determine which images are real or AI-generated? Were you confident in your assessments or were you guessing? Now imagine this same challenge from the perspective of financial institutions. Acquiring new customers and providing financial products are fundamental to their business models, yet they must somehow differentiate between legitimate applicants and sophisticated criminals.
Executing Transactions
Once synthetic accounts are established, generative AI can produce ongoing transactions that closely mirror legitimate customer behavior. Fraudsters program AI tools to mimic normal spending patterns — timing transactions strategically, choosing expected merchant categories, and keeping amounts within anticipated thresholds.
AI systems continually analyze which behaviors trigger fraud alerts and adjust accordingly, creating a dynamic feedback loop. They refine transaction frequency, dollar amounts and merchant choices to stay below detection thresholds. This adaptive behavior erodes the effectiveness of traditional rule-based controls and enables synthetic identities to operate undetected for longer periods.
How Financial Institutions Are Fighting Back
In response, financial institutions are increasingly adopting AI-powered defense mechanisms to identify and prevent synthetic identity fraud. These tools analyze application data for inconsistencies, evaluate government-issued identification for anomalies, and detect signs of AI generation in uploaded documents.
Emerging AI detection systems can identify telltale signs of AI manipulation, such as:
- Pixel-level artifacts in images
- Metadata or software signatures indicating AI-generated documents
- Inconsistencies in audio or video that suggest deepfake manipulation
By integrating these advanced capabilities into onboarding and continuous monitoring processes, financial institutions can better detect fraudulent applications and reduce the risk of synthetic identities entering their systems.
Conclusion and Additional Resources
As AI technology evolves, so do the methods criminals use to exploit it. Generative AI enables unprecedented automation, scale and sophistication in synthetic identity fraud. Financial institutions must continuously reassess and strengthen their controls across the entire customer lifecycle to stay ahead of emerging threats. Proactive evaluation and robust defense strategies are essential to help mitigate the accelerating risks posed by AI-driven fraud.
Read Generative Artificial Intelligence Increases Synthetic Identity Fraud Threats (PDF) and explore the full Synthetic Identity Fraud Mitigation Toolkit for insights and downloadable resources on addressing this type of fraud.
Test your Knowledge Answer Key: Real or AI-Generated Images
House Images: Image A is AI-generated and Image B is a real photo.
Things to consider:
- Do the house and yard appear to be too vibrant?
- Is there a bike in the front yard that is blurry?
- Are the sunlight and shadows consistent?
Office Images: Image A is a real photo and Image B is AI-generated.
Things to consider:
- Do the hands and facial features appear fully formed?
- Is the background blurry due to distance or are details incomplete?
The synthetic identity fraud mitigation toolkit was developed by the Federal Reserve to help educate the industry about synthetic identity fraud and outline potential ways to help detect and mitigate this fraud type. Insights for this toolkit were provided through interviews with industry experts, publicly available research, and team member expertise. This toolkit is not intended to result in any regulatory or reporting requirements, imply any liabilities for fraud loss, or confer any legal status, legal definitions, or legal rights or responsibilities. While use of this toolkit throughout the industry is encouraged, utilization of the toolkit is voluntary at the discretion of each individual entity. Absent written consent, this toolkit may not be used in a manner that suggests the Federal Reserve endorses a third-party product or service.