Methods and systems for enhanced KYC fraud reduction
A system and method for user validation via an identity provider computing device, including a processor and memory device, communicably coupled with a user computing device associated with a user. The identity provider computing device receives a signal associated with a private Fast Identity Online (FIDO) key from the user computing device, verifies that the signal associated with the private FIDO key corresponds to a public FIDO key stored on the memory device, receives data regarding behavioral biometrics sensed by the user computing device, and inputs variables based on the received data into a trained machine learning (ML) model. The ML model outputs a confidence score based on the variables, and the computing device determines if the confidence score satisfies validation criteria. Based on the determination, a notification of whether the user is fraudulent is output to the user computing device and/or a computing system of a service provider.
1 . A system for user validation, the system comprising:
a computing device of an identity provider including at least one processor and at least one memory device, wherein the identity provider computing device is communicably coupled with a user computing device associated with a user, the identity provider computing device configured to:
receive a signal associated with a private Fast Identity Online (FIDO) key from the user computing device;
verify that the signal associated with the private FIDO key corresponds to a public FIDO key stored in the at least one memory device of the identity provider computing device;
receive data regarding behavioral biometrics sensed by the user computing device;
input one or more variables based on the received data into a trained machine learning (ML) model configured to output a confidence score based on the one or more variables, wherein the ML model is at least partially trained on data from in-person know your customer (KYC) onboarding processes and collection of meta data therefrom;
determine if the confidence score satisfies one or more validation criteria; and
based on the determination, output to at least one of the user computing device and a computing system of a service provider a notification of whether the user is fraudulent.
2 . The system of claim 1 , wherein—the determination is that the confidence score does not satisfy the one or more validation criteria, the notification is to the user computing device, the notification includes a request for additional action by or information from the user.
3 . The system of claim 1 , wherein—the determination is that the confidence score does not satisfy the one or more validation criteria, the notification is to the computing system of the service provider, the notification comprises notice that a transaction is declined.
4 . The system of claim 1 , wherein the one or more variables comprise at least two of: speed of navigation (SN), speed of typing (ST), angle of device (DA), malicious process in background (M), and number of previous successful user validations (Ns).
5 . The system of claim 1 , wherein the one or more variables comprise at least two of: device switch from previous session (D), session mode(S), tab switching in session (Ts), size of active window (W), local timestamp of user (T), number of days since last successful user validations (N), and period of inactivity during session (P).
6 . The system of claim 1 , wherein the trained ML model weights each of the one or more variables based on an extent to which each of the one or more variables is indicative of fraudulent activity.
7 . The system of claim 1 , wherein the private FIDO key is associated with biometric authentication from the user via the user computing device.
8 . The system of claim 1 , wherein one or more of the behavioral biometrics are sensed while the user is inputting information into a form on the user computing device.
9 . A user computing device for user validation with a service provider, the user computing device comprising:
a display screen;
an input device configured to receive input from a user;
at least one memory device; and
one or more processors configured and programmed to individually or collectively:
send a signal associated with a private Fast Identity Online (FIDO) key to an identity provider computing device for comparison by the identity provider computing device against a public FIDO key;
receive an indication that the private FIDO key corresponds to the public FIDO key;
provide a fillable form on the display screen in response to receiving an indication that the private FIDO key corresponds to the public FIDO key;
receive behavioral biometric data sensed as the user fills out the fillable form via the input device;
output one or more variables based on the behavioral biometric data to a trained machine learning (ML) model configured to output a confidence score based on the one or more variables; and
display on the display screen a validation determination of whether the user is considered fraudulent based on evaluation of the confidence score using one or more validation criteria.
10 . The user computing device of claim 9 , wherein the validation determination is negative and based on the negative determination the one or more processors are configured and programmed to individually or collectively perform one of the following: (a) display on the display a request for additional action by or information from the user, or (b) send a notification to a computing system of the service provider that a transaction is declined.
11 . The user computing device of claim 9 , wherein the behavioral biometric data or the one or more variables comprise at least two of: speed of navigation (SN), speed of typing (ST), angle of device (DA), malicious process in background (M), number of previous successful user validations (Ns), device switch from previous session (D), session mode(S), tab switching in session (Ts), size of active window (W), local timestamp of user (T), number of days since last successful user validations (N), and period of inactivity during session (P).
12 . The user computing device of claim 9 , wherein the private FIDO key is associated with biometric authentication from the user via the user computing device.
13 . The user computing device of claim 9 , wherein the validation determination is negative and includes an indication from the identity provider computing device that the private FIDO key does not correspond to the public FIDO key.
14 . The user computing device of claim 9 , wherein the one or more processors are configured to receive verification from the identity provider that the private FIDO key corresponds to the public FIDO key, the receipt of the behavioral data from the user being enabled based on the verification of the private FIDO key.
15 . Non-transitory computer-readable storage media having computer-executable instructions for user validation, wherein when executed by at least one processor, the computer-executable instructions cause the at least one processor to:
validate a user computing device via FIDO-based key pairs by verifying that the private FIDO key corresponds to a public FIDO key;
receive from the user computing device details input by a user into a form on the user computing device in response to verifying that the private FIDO key corresponds to the public FIDO key;
receive data regarding behavioral biometrics sensed by the user computing device, wherein at least some of the behavioral biometrics are sensed as the user inputs the details into the form;
determine a confidence score based on the behavioral biometrics data; and
output to at least one of the user computing device and a computing system of a service provider a validation determination of whether the user is fraudulent based on comparison of the confidence score against one or more validation criteria.
16 . The non-transitory computer-readable storage media of claim 15 , wherein the computer-executable instructions cause the at least one processor to input variables based on the behavioral biometrics data into a trained machine learning (ML) model configured to output the confidence score, wherein the trained ML model is trained to weigh each of the variables based on an extent to which each of the variables is indicative of fraudulent activity, wherein the ML model is at least partially trained on data from in-person know your customer (KYC) onboarding processes and collection of meta data therefrom.
17 . The non-transitory computer-readable storage media of claim 15 , wherein the validation determination is that the confidence score does not satisfy the one or more validation criteria, the notification is sent to the user computing device, and the notification includes a request for additional action by or information from the user.
18 . The non-transitory computer-readable storage media of claim 15 , wherein the validation determination is that the confidence score does not satisfy the one or more validation criteria, the notification is sent to the computing device of the service provider, and the notification comprises notice that a transaction is declined.
19 . The non-transitory computer-readable storage media of claim 15 , wherein the one or more variables comprise at least two of: speed of navigation (SN), speed of typing (ST), angle of device (DA), malicious process in background (M), and number of previous successful user validations (Ns).
20 . The non-transitory computer-readable storage media of claim 15 , wherein the one or more variables comprise the number of times tab switching occurs in a form filling session (Ts) and speed of navigation (S N ) being higher or lower than a given deviation from a standard time difference for navigating between given consecutive fields.