Call-Based Fraud Verification for Deepfake Transaction Scams

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

The proliferation of sophisticated digital manipulation techniques in electronic financial transactions, such as deepfakes, enables scammers to impersonate acquaintances and orchestrate fraudulent transactions by exploiting trust and familiarity, necessitating a reliable solution for prevention.

Innovation Solution

A fraud prevention server extracts call content, identifies the caller's identifier, initiates communication with the acquaintance, and instructs a payment application server to reject or hold transactions based on the response, utilizing deepfake detection models and contact list verification to confirm the legitimacy of the call.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If deepfake detection models and call content analysis are used to identify fraudulent calls, then the reliability of transaction verification is improved, but the device complexity and processing time increase

Engineering Contradiction:

Improvetransaction verification reliabilityVSAvoidsystem complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The patent introduces a fraud prevention server as an intermediary component between the communication system and payment processing. This server hosts deepfake detection models and performs call content analysis, acting as a mediator that verifies caller authenticity before transactions proceed. This resolves the contradiction by centralizing complex detection functionality in a dedicated intermediary system rather than distributing it across all devices.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system segments the fraud detection functionality into separate modules: call content extraction, deepfake detection model execution, and verification decision-making. This segmentation allows each component to be optimized independently and reduces overall system complexity by breaking down the monolithic detection process into manageable segments that can be processed sequentially.

Inventive Principle:

Principle #1Segmentation

2Reliability

If real-time call content extraction and deepfake detection are performed, then fraudulent transactions are prevented more effectively, but the processing time and computational resources increase

Engineering Contradiction:

Improvefraud prevention effectivenessVSAvoidtransaction processing time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The system performs preliminary fraud detection by analyzing call content and executing deepfake detection models before the actual financial transaction is authorized. This preliminary action identifies potential fraud early in the process, preventing wasted processing time on fraudulent transactions and allowing legitimate transactions to proceed without delay once verification is complete.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system applies partial detection by focusing deepfake analysis only on specific suspicious patterns or high-risk transaction scenarios rather than performing exhaustive analysis on all calls. This selective approach reduces processing time while maintaining effective fraud prevention for the most critical cases.

Inventive Principle:

Principle #16Partial or excessive action

3Reliability

If the system verifies caller identity through acquaintance confirmation, then the risk of impersonation fraud is reduced, but the ease of operation and user convenience decrease

Engineering Contradiction:

Improveidentity verification accuracyVSAvoidtransaction convenience

Core Design Contradiction:

ReliabilityVSEase of operation

Solution Approach 1:

The system implements self-service verification by automatically contacting the alleged acquaintance through stored contact information and having them confirm or deny the caller's identity without requiring the transaction participant to manually verify. This automated self-service approach maintains high verification accuracy while minimizing the operational burden on users, as the verification process occurs in the background without interrupting the transaction flow.

Inventive Principle:

Principle #25Self-service

Data Source

PatentUS20250390877A1Method and system to prevent fraudulent transactions

Publication Date: 2025.12.25 MASTERCARD INT INC

AI summary

A method for facilitating prevention of fraudulent transactions is provided. A fraud prevention server extracts content of an ongoing call established with a user device of a user by a caller. The fraud prevention server initiates a communication with a communication device of an acquaintance of the user upon determining that a financial transaction is associated with the call and the caller alleges to be the acquaintance of the user based on the extracted content. The communication indicates the acquaintance to deny or confirm whether the ongoing call with the user device is being set-up by the acquaintance. When a response to the communication indicates denial of the ongoing call being set-up by the acquaintance or in an event of absence of a response to the communication, the fraud prevention server notifies a payment application server to reject the financial transaction associated with the caller.