Universal self-service kiosk fraud detection platform

A method, comprising:

• training, by a computing platform, the computing platform having at least one processor and memory and using labelled data from historical transactions indicating whether each transaction of the historical transactions was fraudulent or non-fraudulent, a machine learning model to identify patterns in transaction data to determine whether the transaction data is potentially fraudulent;
• receiving, by the at least one processor, and via a self-service kiosk, a request for transaction;
• determining, by the at least one processor and based on the request for transaction, a financial institution associated with the request for transaction;
• identifying, by the at least one processor, the financial institution as part of a consortium of financial institutions;
• based on identifying the financial institution as part of the consortium of financial institutions, executing, by the at least one processor and in real-time, the machine learning model to identify potential fraud in the request for transaction, wherein executing the machine learning model includes using, as inputs, transaction details from the request for transaction to output a determination of whether potential fraud exists in the request for transaction and wherein the machine learning model is associated with the consortium of financial institutions;
• responsive to determining that potential fraud does not exist in the request for transaction, processing, by the at least one processor, the request for transaction;
• responsive to determining that potential fraud does exist:
• identifying, by the at least one processor, one or more transaction security actions in response to the request for transaction;
• transmitting, by the at least one processor and to the self-service kiosk, the identified one or more transaction security actions, wherein transmitting the identified one or more transaction security actions causes the self-service kiosk to execute the one or more transaction security actions;
• identifying, by the at least one processor, one or more other self-service kiosks within a geo-location;
• identifying, by the at least one processor, one or more self-service kiosk security actions;
• and
• transmitting, by the at least one processor, the one or more self-service kiosk security actions to the one or more other self-service kiosks within the geo-location, wherein transmitting the one or more self-service kiosk security actions to the one or more other self-service kiosks within the geo-location causes the one or more other self-service kiosks to execute the one or more self-service kiosk security actions and wherein executing the one or more self-service kiosk security actions includes disabling the one or more self-service kiosks to prevent further use.

One or more non-transitory computer-readable media storing instructions that, when executed by a computing platform comprising at least one processor, memory, and a communication interface, cause the computing platform to:

• train, using labelled data from historical transactions indicating whether each transaction of the historical transactions was fraudulent or non-fraudulent, a machine learning model to identify patterns in transaction data to determine whether the transaction data is potentially fraudulent;
• receive, via a self-service kiosk, a request for transaction;
• determine, based on the request for transaction, a financial institution associated with the request for transaction;
• identify the financial institution as part of a consortium of financial institutions;
• based on identifying the financial institution as part of the consortium of financial institutions, execute, in real-time, the machine learning model to identify potential fraud in the request for transaction, wherein executing the machine learning model includes using, as inputs, transaction details from the request for transaction to output a determination of whether potential fraud exists in the request for transaction and wherein the machine learning model is associated with the consortium of financial institutions;
• responsive to determining that potential fraud does not exist in the request for transaction, processing the request for transaction;
• responsive to determining that potential fraud does exist:
• identify one or more transaction security actions in response to the request for transaction;
• transmit, to the self-service kiosk, the identified one or more transaction security actions, wherein transmitting the identified one or more transaction security actions causes the self-service kiosk to execute the one or more transaction security actions;
• identify one or more other self-service kiosks within a geo-location;
• identify one or more self-service kiosk security actions;
• and
• transmit the one or more self-service kiosk security actions to the one or more other self-service kiosks within the geo-location, wherein transmitting the one or more self-service kiosk security actions to the one or more other self-service kiosks within the geo-location causes the one or more other self-service kiosks to execute the one or more self-service kiosk security actions and wherein executing the one or more self-service kiosk security actions includes disabling the one or more self-service kiosks to prevent further use.