Platform for identification and exchange of information on security threats and frauds based on monitoring and correaation of various Anti-Fraud IQ communication channels (AFIQ)

The main goal of the project was to develop a system that detects security threats, including real-time fraud and phishing attempts that are popular nowadays, by means of analysis of users’ communications in ICT networks. The main tasks of the system:

• analysis of all currently available communication channels between financial institutions and their customers,
• detection of identity theft,
• warning and prevention of fraud/ extortion,
• alerting on detection of wanted persons and specific events,
• detection of information fraud.

The system we have developed allows to:

• create and compare “profiles” that characterize users, including voice characteristics and patterns of behavior,
• use artificial intelligence algorithms to create a model describing the ways, channels and devices used by the user (including identification of gender, age and language of the speaker, as well as detection of playback, conversion and voice synthesis attacks),
• use a universal API for other ICT applications and in other products of this type.

The results of the project can be useful, among others, in the government sector, banking, and call center systems to detect security threats (real-time fraud attempts and scams).

Michał Hałoń on the Anti-Fraud IQ project:

Given the exponentially growing number of users of online banking, mobile banking and the absolutely widespread use of social networks, one should expect an increasing number of attacks targeting, for example, bank customers. It is clear that only automated anti-fraud systems are able to cope with such an onslaught of information and the increased number of diverse information channels.

The AFIQ project lasted until the end of July 2022 and was funded by the OPIR/NCBR. Biometriq was the project leader, and NASK-PIB was a partner.