Summary

Deception risk detection is a very important capability for certain ICT systems including systems for payment, rental services and credit granting etc.  Traditional deception risk detection mechanisms for these systems are based on service information gathering and rule/pattern matching, etc. As big data and AI technologies evolve, back-end systems with automatic decision-making for deception risk detection are increasingly common. However, the deception risk identification decisions made solely by back-end systems face challenges such as the dependency on adequate data, balance between accuracy and recall rate and a friendly user interaction when service termination caused by deception risk detection and response. Interactive deception risk detection, which means that during the service launching and execution, users are engaged in conversation with ICT service systems, for information queries to identify and reduce deception risk, provides the following benefits:

Firstly, interactive deception risk detection helps gather necessary information during the service process to strengthen the detection and decision of deception risk analysis, which can improve the accuracy and reduce the recall rate.

Secondly, when a suspicious deception risk occurs, interaction with the user helps the user understand what happened and potentially promote the user’s awareness of deception risks in subsequent transactions.

Considering these factors, this document introduces technical capabilities of interactive deception risk detection, with which the deception risk detection system will cooperate with an interactive engine to better identify the deception risk when insufficient information is available to decide the risk level. According to rules, corpus and other contextual factors, the interactive engine contacts user by phone call, SMS, online interaction or other means to collect more information for the deception risk detection system so as to lower the risk level.

This Recommendation specifies technical capabilities of interactive deception risk detection, including capability identification, functional components to support identified capabilities, and procedures that support interaction among different functional components.