Future-Proofing Death Claims: Emerging AI Threats and Tactics

By Margo Jenkins (AAA Life) & Rob Strange (Evadata)
​
The landscape of insurance fraud is rapidly evolving, and life insurance claims are no exception. With the accelerated development and accessibility of Artificial Intelligence (AI) and generative AI tools, claims examiners are facing sophisticated new threats. Understanding these emerging risks and adopting proactive AI-powered defenses are crucial to safeguarding your company’s financial health and maintaining trust with legitimate policyholders.

Two different companies this year have experienced false death certificate submissions:

•  “Familial Death Fraud Claim” Spring 2025 – A life insurance company recently uncovered a fraud scheme involving a falsified death claim submitted by a policy beneficiary. The claim was initiated shortly after the conclusion of a state emergency declaration following a natural disaster that caused widespread flooding. The beneficiary reported the death of the insured and explained that the certified death certificate had been lost in the flood. Instead, she provided a clear photocopy of the certificate, claiming it was the only version they had—saved as an image on their phone. Given the circumstances and the claim amount falling just below the threshold requiring a certified copy, the company accepted the document and processed the payout. Months later, the company received an anonymous tip alleging that the insured was still alive. A subsequent investigation confirmed that the insured and the beneficiary had conspired to submit fraudulent documents to collect the life insurance benefits. This case underscores a critical lesson: vigilant human oversight and robust fraud detection protocols are necessary, especially in scenarios where exceptions are made due to emergencies. It also highlights that, while AI can enhance efficiency, the industry needs AI systems to evolve beyond surface-level document analysis to incorporate contextual risk factors, such as timing, location, and behavioral patterns.
•  “Case of the Not-So-Dead Couple” Summer 2024 – A life & annuity company had a fraudster (unrelated to the husband & wife insureds) trick the Contact Center into sharing information on their policies. The fraudster had both insured personal information and was able to get through the verification process. The bad actor proceeded to file a false death claim and fake death certificate on the wife. The death claim was paid out via ACH to a “mule bank account” with a matching beneficiary name. A month later the fraudster filed a death claim for the husband. While the second claim was in-progress, the husband opened the USPS mailed “Explanation of Benefits” letter and called in to report that “my wife is alive!” This led to an audit of the claim and discovery of red flags that were missed on the death certificate. Fortunately, the carrier was able to freeze the bank account where the funds had been sent on the first death claim. The SIU team was engaged on the second death claim, and it was closed out. It was a bold and creative scheme that almost led to fraud losses on two policies.

 
1. What Are the Threats for False Death Claims?
Historically, false death claims have relied on elaborate schemes and often detectable paper trails. However, generative AI significantly lowers the barrier to entry for fraudsters, enabling the creation of highly convincing, fabricated documents. Here are the primary threats:

• Deepfakes and Synthetic Identities: Criminals can use AI to invent an entirely new, fake person, complete with a seemingly legitimate digital footprint, purchase a life insurance policy, and then stage their "death" to collect the payout.
• AI-Generated Documents: The ability of AI to produce highly polished and seemingly official documents is a game-changer for fraudsters. This includes:
  • Fake Death Certificates: AI can generate official-looking death certificates and coroner's reports that are difficult to distinguish from genuine ones, lacking the typical grammatical errors or inconsistencies often found in amateur forgeries.
  • Fabricated Medical Records: AI can create convincing X-rays, MRI scans, lab results, and other medical documentation to support a fabricated illness, injury, or cause of death. These documents can be tailored to align with a specific narrative for a claim.
• Hybrid Fraud Tactics: Fraudsters are increasingly combining traditional methods with AI-generated content. For example, they might submit one genuine piece of evidence alongside several AI-altered ones to create a more convincing, yet exaggerated, claim. They may also use stolen personal data (from data breaches) in conjunction with deepfakes to bypass identity verification.

​
2. Are False Death Records and Fake Medical Documents a Growing Risk for the Industry?
The risk posed by false death and medical records is accelerating at an alarming rate due to the rapid advancements in generative AI and the ease of access to these tools.

• Exponential Growth in Generative AI Capabilities: The sophistication of AI models for generating realistic text, images, and audio is improving at an exponential pace. What was once easily identifiable as fake just a few years ago can now be nearly indistinguishable from reality. This means the quality and convincing nature of fabricated documents will only continue to improve.
• Dark web vs. Mainstream AI tools: While mainstream AI tools like Google Gemini, ChatGPT, and Co-Pilot will not create an actual death certificate, they will tell you the components of one. Both familial fraudsters and organized fraudsters can access dark web tools. AI tools such as “FraudGPT” and “Chaos-GPT” have been created by hackers. These have no limits or guardrails to prevent creation of fake documents.
• Accessibility and User-Friendliness: AI tools are no longer confined to expert programmers. User-friendly interfaces and readily available platforms allow individuals with minimal technical knowledge to generate sophisticated deepfakes and documents. This widespread accessibility significantly expands the pool of potential fraudsters.

 
3. How Can Life Insurance Carriers Detect Fraud and Prevent It?
Combating AI-powered fraud requires a multi-layered approach that integrates advanced technology with human expertise and continuous adaptation.

• Human Expertise and Training: While technology to detect fraud is evolving, it's not a full substitute for human judgment. Claims examiners must be continuously trained on emerging AI fraud tactics and be equipped with the skills to identify subtle red flags. Special Investigative Units (SIUs) will become even more critical in investigating complex cases flagged by AI. Train your teams on:
  • Using identify verification tools that go beyond “knowledge-based questions” before giving out policy information over the phone.   It’s become too easy for fraudsters to steal or buy personal information.
  • Knowing the common red flags for a death certificate image (see example below).
  • Reviewing the “document properties” of the death certificate or medical documents image file to see if it has been altered or modified (e.g., PDF, JPEG, PNG, GIF).

(a full-size version of death certificate is below for use in fraud training)

• ​Independent Verification of Death: While relying on official death certificates is standard, consider situations where independent verification, such as direct data from the Social Security Death Master File (SSDMF), state vital records, funeral homes, and obituaries can be used for express claims, claims with red flags, or high value claims. There are providers in the industry who give access to this type of data (feel free to contact us for examples).
• Leveraging AI for Fraud Detection (AI vs. AI): This is perhaps the most crucial tactic. Companies must invest in and implement their own AI and machine learning (ML) solutions specifically designed to detect fraudulent patterns and AI-generated content.
  • Deepfake Detection Software: Specialized software can analyze submitted images, videos, and audio for tell-tale signs of AI generation or manipulation (e.g., inconsistent lighting, unnatural movements, or subtle audio artifacts).
  • Voice Analytics and Biometrics: During customer interactions, voice recognition and audio intelligence can analyze a caller's voice for signs of being a recording or a deepfake and even assess stress or inconsistencies that might indicate deception.
• Industry Collaboration and Information Sharing: Fraudsters often target multiple companies. Sharing anonymized fraud data, trends, and detection techniques across the industry can create a collective defense against evolving threats. There are providers in the industry who help carriers collaborate on fraud (feel free to contact us for examples).

While this threat of false death claims and the use of AI tools to support them is an emerging risk, companies need to start taking efforts to detect and prevent it. AI tools are making it easier for both greedy family members and organized fraudsters to submit false claims. The “traveler and community” fraud groups figured out how to get through new business safeguards for taking out coverage, so our industry needs to be prepared for the same thing to happen with death claims.
 

If you have any questions or would like examples of providers who can help with independent verification of death or industry collaboration on fighting fraud, feel free to contact Rob Strange ([email protected]) and Margo Jenkins ([email protected]).

​ 
Appendix: Full-size image of death certificate for use in fraud training on red flags